1253:
using generative models of deep belief nets (DBN) would overcome the main difficulties of neural nets. However, it was discovered that replacing pre-training with large amounts of training data for straightforward backpropagation when using DNNs with large, context-dependent output layers produced error rates dramatically lower than then-state-of-the-art
Gaussian mixture model (GMM)/Hidden Markov Model (HMM) and also than more-advanced generative model-based systems. The nature of the recognition errors produced by the two types of systems was characteristically different, offering technical insights into how to integrate deep learning into the existing highly efficient, run-time speech decoding system deployed by all major speech recognition systems. Analysis around 2009–2010, contrasting the GMM (and other generative speech models) vs. DNN models, stimulated early industrial investment in deep learning for speech recognition. That analysis was done with comparable performance (less than 1.5% in error rate) between discriminative DNNs and generative models. In 2010, researchers extended deep learning from
2188:
3174:, the need for training data does not stop once an ANN is trained. Rather, there is a continued demand for human-generated verification data to constantly calibrate and update the ANN. For this purpose, Facebook introduced the feature that once a user is automatically recognized in an image, they receive a notification. They can choose whether or not they like to be publicly labeled on the image, or tell Facebook that it is not them in the picture. This user interface is a mechanism to generate "a constant stream of verification data" to further train the network in real-time. As Mühlhoff argues, the involvement of human users to generate training and verification data is so typical for most commercial end-user applications of Deep Learning that such systems may be referred to as "human-aided artificial intelligence".
839:(1958) proposed the perceptron, an MLP with 3 layers: an input layer, a hidden layer with randomized weights that did not learn, and an output layer. He later published a 1962 book that also introduced variants and computer experiments, including a version with four-layer perceptrons "with adaptive preterminal networks" where the last two layers have learned weights (here he credits H. D. Block and B. W. Knight). The book cites an earlier network by R. D. Joseph (1960) "functionally equivalent to a variation of" this four-layer system (the book mentions Joseph over 30 times). Should Joseph therefore be considered the originator of proper adaptive
62:
632:, in which a signal may propagate through a layer more than once, the CAP depth is potentially unlimited. No universally agreed-upon threshold of depth divides shallow learning from deep learning, but most researchers agree that deep learning involves CAP depth higher than two. CAP of depth two has been shown to be a universal approximator in the sense that it can emulate any function. Beyond that, more layers do not add to the function approximator ability of the network. Deep models (CAP > two) are able to extract better features than shallow models and hence, extra layers help in learning the features effectively.
1508:
1532:
13320:
6112:
2650:, deep BSDE addresses the computational challenges faced by traditional numerical methods in high-dimensional settings. Specifically, traditional methods like finite difference methods or Monte Carlo simulations often struggle with the curse of dimensionality, where computational cost increases exponentially with the number of dimensions. Deep BSDE methods, however, employ deep neural networks to approximate solutions of high-dimensional partial differential equations (PDEs), effectively reducing the computational burden.
38:
2657:(PINNs) into the deep BSDE framework enhances its capability by embedding the underlying physical laws directly into the neural network architecture. This ensures that the solutions not only fit the data but also adhere to the governing stochastic differential equations. PINNs leverage the power of deep learning while respecting the constraints imposed by the physical models, resulting in more accurate and reliable solutions for financial mathematics problems.
15011:
14991:
3015:
2143:
2822:(ARL) and UT researchers. Deep TAMER used deep learning to provide a robot with the ability to learn new tasks through observation. Using Deep TAMER, a robot learned a task with a human trainer, watching video streams or observing a human perform a task in-person. The robot later practiced the task with the help of some coaching from the trainer, who provided feedback such as "good job" and "bad job".
1270:
2818:(UT) developed a machine learning framework called Training an Agent Manually via Evaluative Reinforcement, or TAMER, which proposed new methods for robots or computer programs to learn how to perform tasks by interacting with a human instructor. First developed as TAMER, a new algorithm called Deep TAMER was later introduced in 2018 during a collaboration between
1793:(computing the gradient on several training examples at once rather than individual examples) speed up computation. Large processing capabilities of many-core architectures (such as GPUs or the Intel Xeon Phi) have produced significant speedups in training, because of the suitability of such processing architectures for the matrix and vector computations.
1119:. These were designed for unsupervised learning of deep generative models. However, those were more computationally expensive compared to backpropagation. Boltzmann machine learning algorithm, published in 1985, was briefly popular before being eclipsed by the backpropagation algorithm in 1986. (p. 112 ). A 1988 network became state of the art in
2751:, well-tuned to their operating environment. A 1995 description stated, "...the infant's brain seems to organize itself under the influence of waves of so-called trophic-factors ... different regions of the brain become connected sequentially, with one layer of tissue maturing before another and so on until the whole brain is mature".
1061:(LSTM), published in 1995. LSTM can learn "very deep learning" tasks with long credit assignment paths that require memories of events that happened thousands of discrete time steps before. That LSTM was not yet the modern architecture, which required a "forget gate", introduced in 1999, which became the standard RNN architecture.
988:(RNN) were further developed in the 1980s. Recurrence is used for sequence processing, and when a recurrent network is unrolled, it mathematically resembles a deep feedforward layer. Consequently, they have similar properties and issues, and their developments had mutual influences. In RNN, two early influential works were the
1800:) is one such kind of neural network. It doesn't require learning rates or randomized initial weights. The training process can be guaranteed to converge in one step with a new batch of data, and the computational complexity of the training algorithm is linear with respect to the number of neurons involved.
1138:(GMM-HMM) technology based on generative models of speech trained discriminatively. Key difficulties have been analyzed, including gradient diminishing and weak temporal correlation structure in neural predictive models. Additional difficulties were the lack of training data and limited computing power.
2888:
In further reference to the idea that artistic sensitivity might be inherent in relatively low levels of the cognitive hierarchy, a published series of graphic representations of the internal states of deep (20-30 layers) neural networks attempting to discern within essentially random data the images
2515:
Deep learning has been shown to produce competitive results in medical application such as cancer cell classification, lesion detection, organ segmentation and image enhancement. Modern deep learning tools demonstrate the high accuracy of detecting various diseases and the helpfulness of their use by
2295:
In 2023 Murray et al. developed a deep learning architecture which was capable of determining whether a defendant should be tried as a child or adult. Their software was able to estimate subject age with significant accuracy. The same team has developed architectures capable of performing ante-mortem
1648:
A deep neural network (DNN) is an artificial neural network with multiple layers between the input and output layers. There are different types of neural networks but they always consist of the same components: neurons, synapses, weights, biases, and functions. These components as a whole function in
1573:
that constitute animal brains. Such systems learn (progressively improve their ability) to do tasks by considering examples, generally without task-specific programming. For example, in image recognition, they might learn to identify images that contain cats by analyzing example images that have been
2844:
A main criticism concerns the lack of theory surrounding some methods. Learning in the most common deep architectures is implemented using well-understood gradient descent. However, the theory surrounding other algorithms, such as contrastive divergence is less clear. (e.g., Does it converge? If so,
2286:
method in which the system "learns from millions of examples". It translates "whole sentences at a time, rather than pieces". Google
Translate supports over one hundred languages. The network encodes the "semantics of the sentence rather than simply memorizing phrase-to-phrase translations". GT uses
2937:
As deep learning moves from the lab into the world, research and experience show that artificial neural networks are vulnerable to hacks and deception. By identifying patterns that these systems use to function, attackers can modify inputs to ANNs in such a way that the ANN finds a match that human
2769:
Although a systematic comparison between the human brain organization and the neuronal encoding in deep networks has not yet been established, several analogies have been reported. For example, the computations performed by deep learning units could be similar to those of actual neurons and neural
1812:
have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer. By 2019, graphics processing units (GPUs), often with AI-specific enhancements, had displaced CPUs as the dominant method for training large-scale
1652:
For example, a DNN that is trained to recognize dog breeds will go over the given image and calculate the probability that the dog in the image is a certain breed. The user can review the results and select which probabilities the network should display (above a certain threshold, etc.) and return
854:
and Lapa in 1965. They regarded it as a form of polynomial regression, or a generalization of
Rosenblatt's perceptron. A 1971 paper described a deep network with eight layers trained by this method, which is based on layer by layer training through regression analysis. Superfluous hidden units are
2945:
In 2016 researchers used one ANN to doctor images in trial and error fashion, identify another's focal points, and thereby generate images that deceived it. The modified images looked no different to human eyes. Another group showed that printouts of doctored images then photographed successfully
1912:
Large-scale automatic speech recognition is the first and most convincing successful case of deep learning. LSTM RNNs can learn "Very Deep
Learning" tasks that involve multi-second intervals containing speech events separated by thousands of discrete time steps, where one time step corresponds to
1856:
are considered promising for energy-efficient deep learning hardware where the same basic device structure is used for both logic operations and data storage. In 2020, Marega et al. published experiments with a large-area active channel material for developing logic-in-memory devices and circuits
1252:
The 2009 NIPS Workshop on Deep
Learning for Speech Recognition was motivated by the limitations of deep generative models of speech, and the possibility that given more capable hardware and large-scale data sets that deep neural nets might become practical. It was believed that pre-training DNNs
10411:
Wu, Yonghui; Schuster, Mike; Chen, Zhifeng; Le, Quoc V; Norouzi, Mohammad; Macherey, Wolfgang; Krikun, Maxim; Cao, Yuan; Gao, Qin; Macherey, Klaus; Klingner, Jeff; Shah, Apurva; Johnson, Melvin; Liu, Xiaobing; Kaiser, Łukasz; Gouws, Stephan; Kato, Yoshikiyo; Kudo, Taku; Kazawa, Hideto; Stevens,
2674:
Traditional weather prediction systems solve a very complex system of partial differential equations. GraphCast is a deep learning based model, trained on a long history of weather data to predict how weather patterns change over time. It is able to predict weather conditions for up to 10 days
2665:
Image reconstruction is the reconstruction of the underlying images from the image-related measurements. Several works showed the better and superior performance of the deep learning methods compared to analytical methods for various applications, e.g., spectral imaging and ultrasound imaging.
1671:
DNNs are typically feedforward networks in which data flows from the input layer to the output layer without looping back. At first, the DNN creates a map of virtual neurons and assigns random numerical values, or "weights", to connections between them. The weights and inputs are multiplied and
2602:
database, offering researchers the opportunity to identify materials with desired properties for various applications. This development has implications for the future of scientific discovery and the integration of AI in material science research, potentially expediting material innovation and
2398:
Recommendation systems have used deep learning to extract meaningful features for a latent factor model for content-based music and journal recommendations. Multi-view deep learning has been applied for learning user preferences from multiple domains. The model uses a hybrid collaborative and
1672:
return an output between 0 and 1. If the network did not accurately recognize a particular pattern, an algorithm would adjust the weights. That way the algorithm can make certain parameters more influential, until it determines the correct mathematical manipulation to fully process the data.
2146:
1639:
As of 2017, neural networks typically have a few thousand to a few million units and millions of connections. Despite this number being several order of magnitude less than the number of neurons on a human brain, these networks can perform many tasks at a level beyond that of humans (e.g.,
7199:
1280:
Although CNNs trained by backpropagation had been around for decades and GPU implementations of NNs for years, including CNNs, faster implementations of CNNs on GPUs were needed to progress on computer vision. Later, as deep learning becomes widespread, specialized hardware and algorithm
2145:
2528:
is always challenging, since many data points must be considered and analyzed before a target segment can be created and used in ad serving by any ad server. Deep learning has been used to interpret large, many-dimensioned advertising datasets. Many data points are collected during the
2150:
2149:
2144:
1777:
regularization randomly omits units from the hidden layers during training. This helps to exclude rare dependencies. Finally, data can be augmented via methods such as cropping and rotating such that smaller training sets can be increased in size to reduce the chances of overfitting.
2151:
1248:
The impact of deep learning in industry began in the early 2000s, when CNNs already processed an estimated 10% to 20% of all the checks written in the US, according to Yann LeCun. Industrial applications of deep learning to large-scale speech recognition started around 2010.
2747:, neural networks employ a hierarchy of layered filters in which each layer considers information from a prior layer (or the operating environment), and then passes its output (and possibly the original input), to other layers. This process yields a self-organizing stack of
601:). The first representational layer may attempt to identify basic shapes such as lines and circles, the second layer may compose and encode arrangements of edges, the third layer may encode a nose and eyes, and the fourth layer may recognize that the image contains a face.
1537:
Subsequent run of the network on an input image (left): The network correctly detects the starfish. However, the weakly weighted association between ringed texture and sea urchin also confers a weak signal to the latter from one of two intermediate nodes. In addition, a
12283:
Lam, Remi; Sanchez-Gonzalez, Alvaro; Willson, Matthew; Wirnsberger, Peter; Fortunato, Meire; Alet, Ferran; Ravuri, Suman; Ewalds, Timo; Eaton-Rosen, Zach; Hu, Weihua; Merose, Alexander; Hoyer, Stephan; Holland, George; Vinyals, Oriol; Stott, Jacklynn (2023-12-22).
11339:
Litjens, Geert; Kooi, Thijs; Bejnordi, Babak
Ehteshami; Setio, Arnaud Arindra Adiyoso; Ciompi, Francesco; Ghafoorian, Mohsen; van der Laak, Jeroen A.W.M.; van Ginneken, Bram; Sánchez, Clara I. (December 2017). "A survey on deep learning in medical image analysis".
2904:
Some deep learning architectures display problematic behaviors, such as confidently classifying unrecognizable images as belonging to a familiar category of ordinary images (2014) and misclassifying minuscule perturbations of correctly classified images (2013).
1932:
language models. This lets the strength of the acoustic modeling aspects of speech recognition be more easily analyzed. The error rates listed below, including these early results and measured as percent phone error rates (PER), have been summarized since 1991.
10412:
Keith; Kurian, George; Patil, Nishant; Wang, Wei; Young, Cliff; Smith, Jason; Riesa, Jason; Rudnick, Alex; Vinyals, Oriol; Corrado, Greg; et al. (2016). "Google's Neural
Machine Translation System: Bridging the Gap between Human and Machine Translation".
1667:
Deep architectures include many variants of a few basic approaches. Each architecture has found success in specific domains. It is not always possible to compare the performance of multiple architectures, unless they have been evaluated on the same data sets.
2734:
in the early 1990s. These developmental theories were instantiated in computational models, making them predecessors of deep learning systems. These developmental models share the property that various proposed learning dynamics in the brain (e.g., a wave of
8298:
Silver, David; Huang, Aja; Maddison, Chris J.; Guez, Arthur; Sifre, Laurent; Driessche, George van den; Schrittwieser, Julian; Antonoglou, Ioannis; Panneershelvam, Veda (January 2016). "Mastering the game of Go with deep neural networks and tree search".
2148:
2593:
by discovering over 2 million new materials within a relatively short timeframe. GNoME employs deep learning techniques to efficiently explore potential material structures, achieving a significant increase in the identification of stable inorganic
7191:
1608:
Typically, neurons are organized in layers. Different layers may perform different kinds of transformations on their inputs. Signals travel from the first (input), to the last (output) layer, possibly after traversing the layers multiple times.
7732:
Fang, Hao; Gupta, Saurabh; Iandola, Forrest; Srivastava, Rupesh; Deng, Li; Dollár, Piotr; Gao, Jianfeng; He, Xiaodong; Mitchell, Margaret; Platt, John C; Lawrence
Zitnick, C; Zweig, Geoffrey (2014). "From Captions to Visual Concepts and Back".
2170:
data set. MNIST is composed of handwritten digits and includes 60,000 training examples and 10,000 test examples. As with TIMIT, its small size lets users test multiple configurations. A comprehensive list of results on this set is available.
642:
Deep learning algorithms can be applied to unsupervised learning tasks. This is an important benefit because unlabeled data are more abundant than the labeled data. Examples of deep structures that can be trained in an unsupervised manner are
3526:
2241:, can be thought of as a representational layer in a deep learning architecture that transforms an atomic word into a positional representation of the word relative to other words in the dataset; the position is represented as a point in a
10764:
1817:
estimated the hardware computation used in the largest deep learning projects from AlexNet (2012) to AlphaZero (2017) and found a 300,000-fold increase in the amount of computation required, with a doubling-time trendline of 3.4 months.
1525:. The starfish match with a ringed texture and a star outline, whereas most sea urchins match with a striped texture and oval shape. However, the instance of a ring textured sea urchin creates a weakly weighted association between them.
2603:
reducing costs in product development. The use of AI and deep learning suggests the possibility of minimizing or eliminating manual lab experiments and allowing scientists to focus more on the design and analysis of unique compounds.
7229:
9751:
Hannun, Awni; Case, Carl; Casper, Jared; Catanzaro, Bryan; Diamos, Greg; Elsen, Erich; Prenger, Ryan; Satheesh, Sanjeev; Sengupta, Shubho; Coates, Adam; Ng, Andrew Y (2014). "Deep Speech: Scaling up end-to-end speech recognition".
2354:
were used for the first time to predict various properties of molecules in a large toxicology data set. In 2019, generative neural networks were used to produce molecules that were validated experimentally all the way into mice.
2938:
observers would not recognize. For example, an attacker can make subtle changes to an image such that the ANN finds a match even though the image looks to a human nothing like the search target. Such manipulation is termed an "
2174:
Deep learning-based image recognition has become "superhuman", producing more accurate results than human contestants. This first occurred in 2011 in recognition of traffic signs, and in 2014, with recognition of human faces.
13257:; Maddison, Chris J.; Guez, Arthur; Sifre, Laurent; Driessche, George van den; Schrittwieser, Julian; Antonoglou, Ioannis; Panneershelvam, Veda; Lanctot, Marc; Dieleman, Sander; Grewe, Dominik; Nham, John; Kalchbrenner, Nal;
1420:(2018) based on the Progressive GAN by Tero Karras et al. Here the GAN generator is grown from small to large scale in a pyramidal fashion. Image generation by GAN reached popular success, and provoked discussions concerning
11027:
1612:
The original goal of the neural network approach was to solve problems in the same way that a human brain would. Over time, attention focused on matching specific mental abilities, leading to deviations from biology such as
2178:
Deep learning-trained vehicles now interpret 360° camera views. Another example is Facial
Dysmorphology Novel Analysis (FDNA) used to analyze cases of human malformation connected to a large database of genetic syndromes.
1601:) between neurons can transmit a signal to another neuron. The receiving (postsynaptic) neuron can process the signal(s) and then signal downstream neurons connected to it. Neurons may have state, generally represented by
10100:
4381:
697:
activation functions and was generalised to feed-forward multi-layer architectures in 1991 by Kurt Hornik. Recent work also showed that universal approximation also holds for non-bounded activation functions such as
2249:(PCFG) implemented by an RNN. Recursive auto-encoders built atop word embeddings can assess sentence similarity and detect paraphrasing. Deep neural architectures provide the best results for constituency parsing,
9204:
7275:
2199:
Closely related to the progress that has been made in image recognition is the increasing application of deep learning techniques to various visual art tasks. DNNs have proven themselves capable, for example, of
10320:
10131:
616:
useful feature representations from the data automatically. This does not eliminate the need for hand-tuning; for example, varying numbers of layers and layer sizes can provide different degrees of abstraction.
2872:, and they are also still a long way from integrating abstract knowledge, such as information about what objects are, what they are for, and how they are typically used. The most powerful A.I. systems, like
1660:. The extra layers enable composition of features from lower layers, potentially modeling complex data with fewer units than a similarly performing shallow network. For instance, it was proved that sparse
6587:
2315:
A large percentage of candidate drugs fail to win regulatory approval. These failures are caused by insufficient efficacy (on-target effect), undesired interactions (off-target effects), or unanticipated
5276:". David E. Rumelhart, James L. McClelland, and the PDP research group. (editors), Parallel distributed processing: Explorations in the microstructure of cognition, Volume 1: Foundation. MIT Press, 1986.
2889:
on which they were trained demonstrate a visual appeal: the original research notice received well over 1,000 comments, and was the subject of what was for a time the most frequently accessed article on
2598:. The system's predictions were validated through autonomous robotic experiments, demonstrating a noteworthy success rate of 71%. The data of newly discovered materials is publicly available through the
448:
into layers and "training" them to process data. The adjective "deep" refers to the use of multiple layers (ranging from three to several hundred or thousands) in the network. Methods used can be either
2693:
that can be used to measure age. Galkin et al. used deep neural networks to train an epigenetic aging clock of unprecedented accuracy using >6,000 blood samples. The clock uses information from 1000
5002:
The Early
Mathematical Manuscripts of Leibniz: Translated from the Latin Texts Published by Carl Immanuel Gerhardt with Critical and Historical Notes (Leibniz published the chain rule in a 1676 memoir)
3567:
6465:
10204:
13593:
7305:
1237:
were developed for generative modeling. They are trained by training one restricted Boltzmann machine, then freezing it and training another one on top of the first one, and so on, then optionally
2245:. Using word embedding as an RNN input layer allows the network to parse sentences and phrases using an effective compositional vector grammar. A compositional vector grammar can be thought of as
855:
pruned using a separate validation set. Since the activation functions of the nodes are Kolmogorov-Gabor polynomials, these were also the first deep networks with multiplicative units or "gates."
10754:
2913:(AGI) architectures. These issues may possibly be addressed by deep learning architectures that internally form states homologous to image-grammar decompositions of observed entities and events.
2147:
1200:(SVMs) became the preferred choices in the 1990s and 2000s, because of artificial neural networks' computational cost and a lack of understanding of how the brain wires its biological networks.
7571:
1578:
as "cat" or "no cat" and using the analytic results to identify cats in other images. They have found most use in applications difficult to express with a traditional computer algorithm using
1307:
achieved for the first time superhuman performance in a visual pattern recognition contest, outperforming traditional methods by a factor of 3. It then won more contests. They also showed how
2770:
populations. Similarly, the representations developed by deep learning models are similar to those measured in the primate visual system both at the single-unit and at the population levels.
2766:, may be closer to biological reality. In this respect, generative neural network models have been related to neurobiological evidence about sampling-based processing in the cerebral cortex.
1605:, typically between 0 and 1. Neurons and synapses may also have a weight that varies as learning proceeds, which can increase or decrease the strength of the signal that it sends downstream.
7221:
7100:
3833:
958:
to apply CNN to phoneme recognition. It used convolutions, weight sharing, and backpropagation. In 1988, Wei Zhang applied a backpropagation-trained CNN to alphabet recognition. In 1989,
13505:
13333:
6409:
Baker, J.; Deng, Li; Glass, Jim; Khudanpur, S.; Lee, C.-H.; Morgan, N.; O'Shaughnessy, D. (2009). "Research Developments and Directions in Speech Recognition and Understanding, Part 1".
2994:
voice command system open a particular web address, and hypothesized that this could "serve as a stepping stone for further attacks (e.g., opening a web page hosting drive-by malware)".
922:
in 1673 to networks of differentiable nodes. The terminology "back-propagating errors" was actually introduced in 1962 by Rosenblatt, but he did not know how to implement this, although
2856:, not as an all-encompassing solution. Despite the power of deep learning methods, they still lack much of the functionality needed to realize this goal entirely. Research psychologist
9174:
6259:
Morgan, Nelson; Bourlard, Hervé; Renals, Steve; Cohen, Michael; Franco, Horacio (1 August 1993). "Hybrid neural network/hidden markov model systems for continuous speech recognition".
10351:
7165:
2758:
algorithm have been proposed in order to increase its processing realism. Other researchers have argued that unsupervised forms of deep learning, such as those based on hierarchical
1874:
for parallel convolutional processing. The authors identify two key advantages of integrated photonics over its electronic counterparts: (1) massively parallel data transfer through
11722:
11573:
De, Shaunak; Maity, Abhishek; Goel, Vritti; Shitole, Sanjay; Bhattacharya, Avik (2017). "Predicting the popularity of instagram posts for a lifestyle magazine using deep learning".
11019:
8757:
10501:
Murray, J., Heng, D., Lygate, A., et al. (2023). "Applying artificial intelligence to determination of legal age of majority from radiographic data". Morphologie. 108 (360): 100723
1849:
has also built a dedicated system to handle large deep learning models, the CS-2, based on the largest processor in the industry, the second-generation Wafer Scale Engine (WSE-2).
1172:
The principle of elevating "raw" features over hand-crafted optimization was first explored successfully in the architecture of deep autoencoder on the "raw" spectrogram or linear
2909:
hypothesized that these behaviors are due to limitations in their internal representations and that these limitations would inhibit integration into heterogeneous multi-component
612:
to transform the data into a more suitable representation for a classification algorithm to operate on. In the deep learning approach, features are not hand-crafted and the model
9002:
10730:
Wallach, Izhar; Dzamba, Michael; Heifets, Abraham (9 October 2015). "AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery".
10092:
4362:
2646:(BSDE). This method is particularly useful for solving high-dimensional problems in financial mathematics. By leveraging the powerful function approximation capabilities of
2046:
The debut of DNNs for speaker recognition in the late 1990s and speech recognition around 2009-2011 and of LSTM around 2003–2007, accelerated progress in eight major areas:
4119:
2754:
A variety of approaches have been used to investigate the plausibility of deep learning models from a neurobiological perspective. On the one hand, several variants of the
1542:
that was not included in the training gives a weak signal for the oval shape, also resulting in a weak signal for the sea urchin output. These weak signals may result in a
8364:
9196:
8971:
14885:
12753:
Testolin, Alberto; Stoianov, Ivilin; Zorzi, Marco (September 2017). "Letter perception emerges from unsupervised deep learning and recycling of natural image features".
1379:
accuracy, known as the "degradation" problem. In 2015, two techniques were developed to train very deep networks: the Highway Network was published in May 2015, and the
10794:
10312:
10123:
7456:
Cireşan, Dan Claudiu; Meier, Ueli; Gambardella, Luca Maria; Schmidhuber, Jürgen (21 September 2010). "Deep, Big, Simple Neural Nets for Handwritten Digit Recognition".
7251:
1771:
1740:
8211:
8154:
13673:
Szegedy, Christian; Zaremba, Wojciech; Sutskever, Ilya; Bruna, Joan; Erhan, Dumitru; Goodfellow, Ian; Fergus, Rob (2013). "Intriguing properties of neural networks".
13474:
3471:
13741:
13562:
13774:
9082:
Viebke, André; Memeti, Suejb; Pllana, Sabri; Abraham, Ajith (2019). "CHAOS: a parallelization scheme for training convolutional neural networks on Intel Xeon Phi".
4228:
6513:
Doddington, G.; Przybocki, M.; Martin, A.; Reynolds, D. (2000). "The NIST speaker recognition evaluation ± Overview, methodology, systems, results, perspective".
620:
The word "deep" in "deep learning" refers to the number of layers through which the data is transformed. More precisely, deep learning systems have a substantial
13372:
9993:
9305:
P, JouppiNorman; YoungCliff; PatilNishant; PattersonDavid; AgrawalGaurav; BajwaRaminder; BatesSarah; BhatiaSuresh; BodenNan; BorchersAl; BoyleRick (2017-06-24).
3439:
13994:
13843:
12643:
O'Reilly, Randall C. (1 July 1996). "Biologically Plausible Error-Driven Learning Using Local Activation Differences: The Generalized Recirculation Algorithm".
10648:
9729:
8081:
11622:
8571:
2950:
that can then find other instances of it. A refinement is to search using only parts of the image, to identify images from which that piece may have been taken
10444:
9373:
3759:
1886:, and (2) extremely high data modulation speeds. Their system can execute trillions of multiply-accumulate operations per second, indicating the potential of
10389:
7801:
He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian (2016). "Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification".
6543:
Heck, L.; Konig, Y.; Sonmez, M.; Weintraub, M. (2000). "Robustness to Telephone Handset Distortion in Speaker Recognition by Discriminative Feature Design".
2639:
321:
10704:
10196:
6385:
5727:
13585:
10889:
8671:
5906:(2020). "Generative Adversarial Networks are Special Cases of Artificial Curiosity (1990) and also Closely Related to Predictability Minimization (1991)".
1656:
DNNs can model complex non-linear relationships. DNN architectures generate compositional models where the object is expressed as a layered composition of
10941:
10910:
Tkachenko, Yegor (8 April 2015). "Autonomous CRM Control via CLV Approximation with Deep Reinforcement Learning in Discrete and Continuous Action Space".
7297:
6762:
3829:(1986). Learning while searching in constraint-satisfaction problems. University of California, Computer Science Department, Cognitive Systems Laboratory.
2187:
539:. However, current neural networks do not intend to model the brain function of organisms, and are generally seen as low-quality models for that purpose.
11915:"Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations"
11401:
Forslid, Gustav; Wieslander, Hakan; Bengtsson, Ewert; Wahlby, Carolina; Hirsch, Jan-Michael; Stark, Christina Runow; Sadanandan, Sajith Kecheril (2017).
7561:
Ciresan, Dan; Giusti, Alessandro; Gambardella, Luca M.; Schmidhuber, Jürgen (2012). Pereira, F.; Burges, C. J. C.; Bottou, L.; Weinberger, K. Q. (eds.).
2228:
Neural networks have been used for implementing language models since the early 2000s. LSTM helped to improve machine translation and language modeling.
1916:
The initial success in speech recognition was based on small-scale recognition tasks based on TIMIT. The data set contains 630 speakers from eight major
11228:
10931:
van den Oord, Aaron; Dieleman, Sander; Schrauwen, Benjamin (2013). Burges, C. J. C.; Bottou, L.; Welling, M.; Ghahramani, Z.; Weinberger, K. Q. (eds.).
10816:
Gilmer, Justin; Schoenholz, Samuel S.; Riley, Patrick F.; Vinyals, Oriol; Dahl, George E. (2017-06-12). "Neural Message Passing for Quantum Chemistry".
9235:
8393:
6486:
Deng, L.; Hassanein, K.; Elmasry, M. (1994). "Analysis of correlation structure for a neural predictive model with applications to speech recognition".
1284:
A key advance for the deep learning revolution was hardware advances, especially GPU. Some early work dated back to 2004. In 2009, Raina, Madhavan, and
833:
produced work on "Intelligent Machinery" that was not published in his lifetime, containing "ideas related to artificial evolution and learning RNNs."
713:
concerns the capacity of networks with bounded width but the depth is allowed to grow. Lu et al. proved that if the width of a deep neural network with
624:(CAP) depth. The CAP is the chain of transformations from input to output. CAPs describe potentially causal connections between input and output. For a
14133:
5998:
13443:
10261:
Kariampuzha, William; Alyea, Gioconda; Qu, Sue; Sanjak, Jaleal; Mathé, Ewy; Sid, Eric; Chatelaine, Haley; Yadaw, Arjun; Xu, Yanji; Zhu, Qian (2023).
8178:
Li, Xiangang; Wu, Xihong (2014). "Constructing Long Short-Term Memory based Deep Recurrent Neural Networks for Large Vocabulary Speech Recognition".
8102:
Singh, Premjeet; Saha, Goutam; Sahidullah, Md (2021). "Non-linear frequency warping using constant-Q transformation for speech emotion recognition".
7595:
Ciresan, D.; Giusti, A.; Gambardella, L.M.; Schmidhuber, J. (2013). "Mitosis Detection in Breast Cancer Histology Images with Deep Neural Networks".
7562:
4956:
Fukushima, K. (1980). "Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position".
1412:'s principle of artificial curiosity) became state of the art in generative modeling during 2014-2018 period. Excellent image quality is achieved by
412:
13882:
2876:(...) use techniques like deep learning as just one element in a very complicated ensemble of techniques, ranging from the statistical technique of
1653:
the proposed label. Each mathematical manipulation as such is considered a layer, and complex DNN have many layers, hence the name "deep" networks.
1466:), as well as a range of large-vocabulary speech recognition tasks have steadily improved. Convolutional neural networks were superseded for ASR by
14727:
13652:
Nguyen, Anh; Yosinski, Jason; Clune, Jeff (2014). "Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images".
9802:
Cireşan, Dan; Meier, Ueli; Masci, Jonathan; Schmidhuber, Jürgen (August 2012). "Multi-column deep neural network for traffic sign classification".
9148:
8284:
13171:
8939:
7093:
3830:
2428:
In medical informatics, deep learning was used to predict sleep quality based on data from wearables and predictions of health complications from
13497:
10226:
Brocardo, Marcelo Luiz; Traore, Issa; Woungang, Isaac; Obaidat, Mohammad S. (2017). "Authorship verification using deep belief network systems".
1375:
In 2014, the state of the art was training “very deep neural network” with 20 to 30 layers. Stacking too many layers led to a steep reduction in
938:
applied backpropagation to neural networks in 1982 (his 1974 PhD thesis, reprinted in a 1994 book, did not yet describe the algorithm). In 1986,
13341:
11891:
10560:
Verbist, B; Klambauer, G; Vervoort, L; Talloen, W; The Qstar, Consortium; Shkedy, Z; Thas, O; Bender, A; Göhlmann, H. W.; Hochreiter, S (2015).
7978:
Karras, T.; Aila, T.; Laine, S.; Lehtinen, J. (26 February 2018). "Progressive Growing of GANs for Improved Quality, Stability, and Variation".
5128:
3218:
2557:. These applications include learning methods such as "Shrinkage Fields for Effective Image Restoration" which trains on an image dataset, and
1166:
1011:
where each RNN tries to predict its own next input, which is the next unexpected input of the RNN below. This "neural history compressor" uses
11049:
Chicco, Davide; Sadowski, Peter; Baldi, Pierre (1 January 2014). "Deep autoencoder neural networks for gene ontology annotation predictions".
8003:
7927:
2964:
into thinking ordinary people were celebrities, potentially allowing one person to impersonate another. In 2017 researchers added stickers to
2864:
Realistically, deep learning is only part of the larger challenge of building intelligent machines. Such techniques lack ways of representing
585:
in which a hierarchy of layers is used to transform input data into a slightly more abstract and composite representation. For example, in an
13804:
12688:"Probabilistic Models and Generative Neural Networks: Towards an Unified Framework for Modeling Normal and Impaired Neurocognitive Functions"
10470:
9166:
1169:
Speaker Recognition benchmark. It was deployed in the Nuance Verifier, representing the first major industrial application of deep learning.
11744:
10343:
10157:; He, X.; Heck, L.; Tur, G.; Yu, D.; Zweig, G. (2015). "Using recurrent neural networks for slot filling in spoken language understanding".
7917:
Goodfellow, Ian; Pouget-Abadie, Jean; Mirza, Mehdi; Xu, Bing; Warde-Farley, David; Ozair, Sherjil; Courville, Aaron; Bengio, Yoshua (2014).
7157:
5111:
Ostrovski, G.M., Volin,Y.M., and Boris, W.W. (1971). On the computation of derivatives. Wiss. Z. Tech. Hochschule for Chemistry, 13:382–384.
1707:
DNNs are prone to overfitting because of the added layers of abstraction, which allow them to model rare dependencies in the training data.
628:, the depth of the CAPs is that of the network and is the number of hidden layers plus one (as the output layer is also parameterized). For
7755:
Kiros, Ryan; Salakhutdinov, Ruslan; Zemel, Richard S (2014). "Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models".
7192:"Keynote talk: 'Achievements and Challenges of Deep Learning - From Speech Analysis and Recognition To Language and Multimodal Processing'"
3963:
2946:
tricked an image classification system. One defense is reverse image search, in which a possible fake image is submitted to a site such as
2210: – capturing the style of a given artwork and applying it in a visually pleasing manner to an arbitrary photograph or video
1084:
to predict the reactions of the environment to these patterns. This was called "artificial curiosity". In 2014, this principle was used in
982:
et al., that classifies digits, was applied by several banks to recognize hand-written numbers on checks digitized in 32x32 pixel images.
11714:
11516:"System for the Recognizing of Pigmented Skin Lesions with Fusion and Analysis of Heterogeneous Data Based on a Multimodal Neural Network"
3890:
Co-evolving recurrent neurons learn deep memory POMDPs. Proc. GECCO, Washington, D. C., pp. 1795–1802, ACM Press, New York, NY, USA, 2005.
11253:
4637:
Unpublished (Later Published in Ince DC, Editor, Collected Works of AM Turing—Mechanical Intelligence, Elsevier Science Publishers, 1992)
1796:
Alternatively, engineers may look for other types of neural networks with more straightforward and convergent training algorithms. CMAC (
874:
to classify non-linearily separable pattern classes. Subsequent developments in hardware and hyperparameter tunings have made end-to-end
208:
173:
1513:
Simplified example of training a neural network in object detection: The network is trained by multiple images that are known to depict
1257:
to large vocabulary speech recognition, by adopting large output layers of the DNN based on context-dependent HMM states constructed by
1003:
In the 1980s, backpropagation did not work well for deep learning with long credit assignment paths. To overcome this problem, in 1991,
9696:
9460:
Feldmann, J.; Youngblood, N.; Karpov, M.; et al. (2021). "Parallel convolutional processing using an integrated photonic tensor".
5169:
1548:
In reality, textures and outlines would not be represented by single nodes, but rather by associated weight patterns of multiple nodes.
13931:"Human-aided artificial intelligence: Or, how to run large computations in human brains? Toward a media sociology of machine learning"
10967:"The Deep Learning–Based Recommender System "Pubmender" for Choosing a Biomedical Publication Venue: Development and Validation Study"
9867:
8445:
7044:; Kingsbury, B. (2012). "Deep Neural Networks for Acoustic Modeling in Speech Recognition: The Shared Views of Four Research Groups".
3111:
Most Deep Learning systems rely on training and verification data that is generated and/or annotated by humans. It has been argued in
7296:
Deng, Li; Li, Jinyu; Huang, Jui-Ting; Yao, Kaisheng; Yu, Dong; Seide, Frank; Seltzer, Mike; Zweig, Geoff; He, Xiaodong (1 May 2013).
6857:
2643:
1924:, where each speaker reads 10 sentences. Its small size lets many configurations be tried. More importantly, the TIMIT task concerns
10683:
11678:
Kleanthous, Christos; Chatzis, Sotirios (2020). "Gated Mixture Variational Autoencoders for Value Added Tax audit case selection".
11654:
11436:
10066:
8598:
Jozefowicz, Rafal; Vinyals, Oriol; Schuster, Mike; Shazeer, Noam; Wu, Yonghui (2016). "Exploring the Limits of Language Modeling".
5269:
3208:
2253:, information retrieval, spoken language understanding, machine translation, contextual entity linking, writing style recognition,
1797:
1376:
272:
250:
4917:
Fukushima, K. (1979). "Neural network model for a mechanism of pattern recognition unaffected by shift in position—Neocognitron".
14243:
12352:
9398:
Marega, Guilherme Migliato; Zhao, Yanfei; Avsar, Ahmet; Wang, Zhenyu; Tripati, Mukesh; Radenovic, Aleksandra; Kis, Anras (2020).
9349:
5456:"Computerized detection of clustered microcalcifications in digital mammograms using a shift-invariant artificial neural network"
4814:
4116:
3183:
2529:
request/serve/click internet advertising cycle. This information can form the basis of machine learning to improve ad selection.
186:
11280:
10030:
8358:
6829:
14126:
8963:
6621:
1st Intl. Workshop on Biologically Inspired Approaches to Advanced Information Technology, Bio-ADIT 2004, Lausanne, Switzerland
5000:
110:
7537:
1019:
at multiple self-organizing time scales. This can substantially facilitate downstream deep learning. The RNN hierarchy can be
639:
layer-by-layer method. Deep learning helps to disentangle these abstractions and pick out which features improve performance.
14029:
13730:
Miller, G. A., and N. Chomsky. "Pattern conception". Paper for Conference on pattern detection, University of Michigan. 1957.
13411:
12459:
11590:
10786:
8903:
8842:
8201:"Unidirectional Long Short-Term Memory Recurrent Neural Network with Recurrent Output Layer for Low-Latency Speech Synthesis"
8129:
7880:
7612:
7440:
7141:
6683:
6196:
5761:
4744:
4596:
4375:
4270:
4097:
4067:
3955:
3868:
3733:
3400:
2586:
2487:' color values to probabilities over possible image classes. In practice, the probability distribution of Y is obtained by a
1322:
created an FNN that learned to recognize higher-level concepts, such as cats, only from watching unlabeled images taken from
1208:
405:
331:
285:
240:
235:
11090:
9961:
Goldberg, Yoav; Levy, Omar (2014). "word2vec Explained: Deriving Mikolov et al.'s Negative-Sampling Word-Embedding Method".
8200:
8161:
8061:
7679:
6624:
6316:
3079:
14916:
14076:
13466:
11768:
Merchant, Amil; Batzner, Simon; Schoenholz, Samuel S.; Aykol, Muratahan; Cheon, Gowoon; Cubuk, Ekin Dogus (December 2023).
10971:
8542:
7710:
Vinyals, Oriol; Toshev, Alexander; Bengio, Samy; Erhan, Dumitru (2014). "Show and Tell: A Neural Image Caption Generator".
3463:
3188:
2623:
in both forward and inverse problems in a data driven manner. One example is the reconstructing fluid flow governed by the
2443:, a deep-learning based system, achieved a level of accuracy significantly higher than all previous computational methods.
2156:
1241:
using supervised backpropagation. They could model high-dimensional probability distributions, such as the distribution of
485:
17:
13745:
13558:
12987:
Yamins, Daniel L K; DiCarlo, James J (March 2016). "Using goal-driven deep learning models to understand sensory cortex".
7398:; Chen, Yu-Hsin; Yang, Tien-Ju; Emer, Joel (2017). "Efficient Processing of Deep Neural Networks: A Tutorial and Survey".
7375:
3725:
Human Behavior and Another Kind in Consciousness: Emerging Research and Opportunities: Emerging Research and Opportunities
3051:
3001:", false data is continually smuggled into a machine learning system's training set to prevent it from achieving mastery.
2998:
15017:
14568:
14305:
13766:
2839:
1035:
network. In 1993, a neural history compressor solved a "Very Deep Learning" task that required more than 1000 subsequent
790:
384:
356:
351:
245:
9030:
Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis on - SC '17
8515:
7926:. Proceedings of the International Conference on Neural Information Processing Systems (NIPS 2014). pp. 2672–2680.
4205:
1789:
for optimal parameters may not be feasible due to the cost in time and computational resources. Various tricks, such as
12475:
Shrager, J.; Johnson, MH (1996). "Dynamic plasticity influences the emergence of function in a simple cortical array".
9555:
9060:
8622:
Gillick, Dan; Brunk, Cliff; Vinyals, Oriol; Subramanya, Amarnag (2015). "Multilingual Language Processing from Bytes".
8268:
6569:
L.P Heck and R. Teunen. "Secure and Convenient Transactions with Nuance Verifier". Nuance Users Conference, April 1998.
6294:
3428:
2246:
1829:
were designed to speed up deep learning algorithms. Deep learning processors include neural processing units (NPUs) in
721:; if the width is smaller or equal to the input dimension, then a deep neural network is not a universal approximator.
344:
213:
203:
193:
13364:
9986:
5061:
The representation of the cumulative rounding error of an algorithm as a Taylor expansion of the local rounding errors
1488:
for "conceptual and engineering breakthroughs that have made deep neural networks a critical component of computing."
1273:
How deep learning is a subset of machine learning and how machine learning is a subset of artificial intelligence (AI)
14829:
14456:
14263:
14119:
14048:
13978:
13628:
11422:
11076:
10640:
10562:"Using transcriptomics to guide lead optimization in drug discovery projects: Lessons learned from the QSTAR project"
9721:
9532:
9265:
9046:
8866:
Bengio, Yoshua; Boulanger-Lewandowski, Nicolas; Pascanu, Razvan (2013). "Advances in optimizing recurrent networks".
8423:
8077:
7952:
6655:(2006). "Connectionist temporal classification: Labelling unsegmented sequence data with recurrent neural networks".
6310:
6034:
5825:
5325:
5203:
5010:
3098:
3058:
2283:
2279:
1786:
974:
hardware. In 1991, a CNN was applied to medical image object segmentation and breast cancer detection in mammograms.
548:
433:
316:
262:
228:
95:
13835:
11614:
10619:
8060:
Google Research Blog. The neural networks behind Google Voice transcription. August 11, 2015. By Françoise Beaufays
4316:
Sonoda, Sho; Murata, Noboru (2017). "Neural network with unbounded activation functions is universal approximator".
1781:
DNNs must consider many training parameters, such as the size (number of layers and number of units per layer), the
843:
with learning hidden units? Unfortunately, the learning algorithm was not a functional one, and fell into oblivion.
14784:
10434:
6648:
3750:
2654:
2468:
2364:
1204:
870:. In computer experiments conducted by Amari's student Saito, a five layer MLP with two modifiable layers learned
749:
398:
302:
148:
10381:
7019:
4464:
Amari, Shun-Ichi (1972). "Learning patterns and pattern sequences by self-organizing nets of threshold elements".
4408:
2830:
Deep learning has attracted both criticism and comment, in some cases from outside the field of computer science.
989:
10708:
7094:"New types of deep neural network learning for speech recognition and related applications: An overview (ICASSP)"
5723:
3263:
2910:
2500:
993:
671:
80:
10881:
8644:
7896:
Gatys, Leon A.; Ecker, Alexander S.; Bethge, Matthias (26 August 2015). "A Neural Algorithm of Artistic Style".
6338:
5455:
5408:
5362:"Parallel distributed processing model with local space-invariant interconnections and its optical architecture"
5361:
2627:. Using physics informed neural networks does not require the often expensive mesh generation that conventional
2611:
The United States Department of Defense applied deep learning to train robots in new tasks through observation.
528:
programs, where they have produced results comparable to and in some cases surpassing human expert performance.
14971:
14911:
14509:
14103:
10932:
9781:
9558:(30 September 1991). "Several Improvements to a Recurrent Error Propagation Network Phone Recognition System".
7657:
Simonyan, Karen; Andrew, Zisserman (2014). "Very Deep Convolution Networks for Large Scale Image Recognition".
6718:
3065:
3036:
3032:
2853:
2675:
globally, at a very detailed level, and in under a minute, with precision similar to state of the art systems.
2455:
and called Neural Joint Entropy Estimator (NJEE). Such an estimation provides insights on the effects of input
1439:
In 2015, Google's speech recognition improved by 49% by an LSTM-based model, which they made available through
1401:
1319:
1130:
have been explored for many years. These methods never outperformed non-uniform internal-handcrafting Gaussian
1085:
748:, respectively. More specifically, the probabilistic interpretation considers the activation nonlinearity as a
481:
14082:
13105:"Deep Neural Networks Reveal a Gradient in the Complexity of Neural Representations across the Ventral Stream"
9610:
8385:
5808:
Gers, Felix; Schmidhuber, Jürgen; Cummins, Fred (1999). "Learning to forget: Continual prediction with LSTM".
3373:
Ciresan, D.; Meier, U.; Schmidhuber, J. (2012). "Multi-column deep neural networks for image classification".
2088:
Other types of deep models including tensor-based models and integrated deep generative/discriminative models.
531:
Early forms of neural networks were inspired by information processing and distributed communication nodes in
14504:
14193:
11459:"Liver Cancer Detection Using Hybridized Fully Convolutional Neural Network Based on Deep Learning Framework"
10313:"Deep Learning for Natural Language Processing: Theory and Practice (CIKM2014 Tutorial) - Microsoft Research"
9227:
5657:"Learning complex, extended sequences using the principle of history compression (based on TR FKI-148, 1991)"
2975:, potentially leading attackers and defenders into an arms race similar to the kind that already defines the
2922:
2819:
2815:
2620:
2464:
1349:
by a significant margin over shallow machine learning methods. Further incremental improvements included the
847:
477:
3135:(the embedding of annotation or computation tasks in the flow of a game), (2) "trapping and tracking" (e.g.
817:
which is essentially a non-learning RNN architecture consisting of neuron-like threshold elements. In 1972,
14946:
14343:
14300:
14253:
14248:
13435:
12510:
Quartz, SR; Sejnowski, TJ (1997). "The neural basis of cognitive development: A constructivist manifesto".
9134:
Ting Qin, et al. "A learning algorithm of CMAC based on RLS". Neural Processing Letters 19.1 (2004): 49-61.
6796:
5314:
IEEE Transactions on Acoustics, Speech, and Signal Processing, Volume 37, No. 3, pp. 328. – 339 March 1989.
2939:
2628:
2436:
2072:
1708:
1685:
1451:
1450:
Deep learning is part of state-of-the-art systems in various disciplines, particularly computer vision and
1120:
1108:
897:
757:
556:
552:
13866:
13334:"A Google DeepMind Algorithm Uses Deep Learning and More to Master the Game of Go | MIT Technology Review"
12806:"Neural Dynamics as Sampling: A Model for Stochastic Computation in Recurrent Networks of Spiking Neurons"
5310:
4896:
Ramachandran, Prajit; Barret, Zoph; Quoc, V. Le (October 16, 2017). "Searching for Activation Functions".
4289:
Fukushima, K. (1969). "Visual feature extraction by a multilayered network of analog threshold elements".
3047:
14997:
14293:
14219:
13394:
Bradley Knox, W.; Stone, Peter (2008). "TAMER: Training an Agent Manually via Evaluative Reinforcement".
10837:
Zhavoronkov, Alex (2019). "Deep learning enables rapid identification of potent DDR1 kinase inhibitors".
5196:
The Roots of Backpropagation : From Ordered Derivatives to Neural Networks and Political Forecasting
3601:
Bengio, Y.; Courville, A.; Vincent, P. (2013). "Representation Learning: A Review and New Perspectives".
2624:
2369:
2223:
1913:
about 10 ms. LSTM with forget gates is competitive with traditional speech recognizers on certain tasks.
1887:
1617:, or passing information in the reverse direction and adjusting the network to reflect that information.
1238:
1091:
During 1985–1995, inspired by statistical mechanics, several architectures and methods were developed by
886:
875:
863:
682:
501:
267:
218:
115:
13163:
9513:
Garofolo, J.S.; Lamel, L.F.; Fisher, W.M.; Fiscus, J.G.; Pallett, D.S.; Dahlgren, N.L.; Zue, V. (1993).
9144:
7040:
Hinton, G.; Deng, L.; Yu, D.; Dahl, G.; Mohamed, A.; Jaitly, N.; Senior, A.; Vanhoucke, V.; Nguyen, P.;
6703:, in Bengio, Yoshua; Schuurmans, Dale; Lafferty, John; Williams, Chris K. I.; and Culotta, Aron (eds.),
5295:. Meeting of the Institute of Electrical, Information and Communication Engineers (IEICE). Tokyo, Japan.
14621:
14556:
14157:
13536:
8928:
7636:
Ng, Andrew; Dean, Jeff (2012). "Building High-level Features Using Large Scale Unsupervised Learning".
7222:"Roles of Pre-Training and Fine-Tuning in Context-Dependent DBN-HMMs for Real-World Speech Recognition"
6022:
Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Volume 1: Foundations
5785:
3228:
3198:
3140:
2782:
2699:
1369:
1292:
GPUs, an early demonstration of GPU-based deep learning. They reported up to 70 times faster training.
1192:
Neural networks entered a null, and simpler models that use task-specific handcrafted features such as
1050:
794:
658:
in 1986, and to artificial neural networks by Igor Aizenberg and colleagues in 2000, in the context of
625:
90:
73:
31:
11052:
Proceedings of the 5th ACM Conference on Bioinformatics, Computational Biology, and Health Informatics
6705:
Advances in Neural Information Processing Systems 22 (NIPS'22), December 7th–10th, 2009, Vancouver, BC
6613:
6580:"Acoustic Modeling with Deep Neural Networks Using Raw Time Signal for LVCSR (PDF Download Available)"
5960:
4138:"Efficient probabilistic inference in generic neural networks trained with non-probabilistic feedback"
15022:
14880:
14519:
14350:
14173:
13250:
11883:
11514:
Lyakhov, Pavel Alekseevich; Lyakhova, Ulyana Alekseevna; Nagornov, Nikolay Nikolaevich (2022-04-03).
5847:(1991). "A possibility for implementing curiosity and boredom in model-building neural controllers".
5132:
4720:
Contributions to Perceptron Theory, Cornell Aeronautical Laboratory Report No. VG-11 96--G-7, Buffalo
3243:
2961:
1774:
1675:
1570:
1522:
1497:
951:
919:
802:
753:
594:
473:
168:
13708:
13319:
9816:
8886:
7918:
6111:
6013:
5708:
Page 150 ff demonstrates credit assignment across the equivalent of 1,200 layers in an unfolded RNN.
3546:
1180:
features that contain stages of fixed transformation from spectrograms. The raw features of speech,
1049:'s diploma thesis (1991) implemented the neural history compressor, and identified and analyzed the
61:
14921:
14178:
12524:
9288:"HUAWEI Reveals the Future of Mobile AI at IFA 2017 | HUAWEI Latest News | HUAWEI Global"
9287:
8076:
Sak, Haşim; Senior, Andrew; Rao, Kanishka; Beaufays, Françoise; Schalkwyk, Johan (September 2015).
7999:
6665:
5522:
2731:
2476:
2429:
2254:
2082:
1150:
1077:
1016:
1008:
985:
871:
826:
629:
454:
292:
13796:
11575:
2017 2nd International Conference on Communication Systems, Computing and IT Applications (CSCITA)
10477:
14966:
14951:
14604:
14599:
14499:
14367:
14148:
12220:"Training Variational Networks With Multidomain Simulations: Speed-of-Sound Image Reconstruction"
10122:
Huang, Po-Sen; He, Xiaodong; Gao, Jianfeng; Deng, Li; Acero, Alex; Heck, Larry (1 October 2013).
8155:"Long Short-Term Memory recurrent neural network architectures for large scale acoustic modeling"
6879:
5584:
3131:
distinguishes five types of "machinic capture" of human microwork to generate training data: (1)
3025:
2926:
2703:
2381:
1861:
1826:
1661:
1380:
1216:
1054:
967:
613:
465:
437:
53:
10965:
Feng, X.Y.; Zhang, H.; Ren, Y.J.; Shang, P.H.; Zhu, Y.; Liang, Y.C.; Guan, R.C.; Xu, D. (2019).
10054:
Socher, R.; Perelygin, A.; Wu, J.; Chuang, J.; Manning, C.D.; Ng, A.; Potts, C. (October 2013).
4761:
3493:
1649:
a way that mimics functions of the human brain, and can be trained like any other ML algorithm.
1386:
Around the same time, deep learning started impacting the field of art. Early examples included
1203:
In 2003, LSTM became competitive with traditional speech recognizers on certain tasks. In 2006,
604:
Importantly, a deep learning process can learn which features to optimally place at which level
15045:
14926:
14686:
14405:
14400:
13703:
12519:
9811:
8881:
6660:
5517:
3541:
3120:
2811:
In 2017, Covariant.ai was launched, which focuses on integrating deep learning into factories.
2275:
1838:
1749:
1718:
1579:
1300:
1197:
1177:
1058:
1024:
840:
517:
163:
11020:"A Multi-View Deep Learning Approach for Cross Domain User Modeling in Recommendation Systems"
6026:
2589:
announced that they had developed an AI system known as GNoME. This system has contributed to
14956:
14941:
14906:
14594:
14494:
14362:
9872:
5992:
4652:"The perceptron: A probabilistic model for information storage and organization in the brain"
4087:
3723:
2918:
2789:
2207:
2125:
1842:
1507:
1394:(2015), both of which were based on pretrained image classification neural networks, such as
1391:
1036:
859:
798:
760:
in neural networks. The probabilistic interpretation was introduced by researchers including
717:
activation is strictly larger than the input dimension, then the network can approximate any
703:
560:
489:
458:
14824:
9663:
7533:
6652:
5903:
5862:
5844:
5781:
5627:
5158:
4496:
1409:
1304:
1065:
1004:
14976:
14931:
14377:
14322:
14168:
14163:
13276:
13206:
12817:
12589:
12307:
12163:"High-Resolution Multi-Spectral Imaging With Diffractive Lenses and Learned Reconstruction"
12115:
12048:
11991:
11926:
11781:
11470:
11359:
9411:
9101:
8563:
8453:
8308:
7339:
7053:
6977:
6418:
6145:
6136:; Neal, Radford (1995-05-26). "The wake-sleep algorithm for unsupervised neural networks".
5467:
5420:
5373:
5232:
4533:
4435:
4149:
3972:
3850:
3799:
3328:
3213:
3072:
2917:(visual or linguistic) from training data would be equivalent to restricting the system to
2647:
2504:
2480:
2472:
2377:
2351:
2339:. AtomNet was used to predict novel candidate biomolecules for disease targets such as the
1871:
1633:
1463:
1346:
1145:
researchers moved away from neural nets to pursue generative modeling. An exception was at
1116:
997:
718:
686:
105:
10154:
6918:
6853:
2983:
software by repeatedly attacking a defense with malware that was continually altered by a
662:
threshold neurons. Although the history of its appearance is apparently more complicated.
8:
14551:
14529:
14278:
14273:
14231:
14183:
13265:(28 January 2016). "Mastering the game of Go with deep neural networks and tree search".
11402:
10679:
10091:
Shen, Yelong; He, Xiaodong; Gao, Jianfeng; Deng, Li; Mesnil, Gregoire (1 November 2014).
8822:
8062:
http://googleresearch.blogspot.co.at/2015/08/the-neural-networks-behind-google-voice.html
8025:
Sohl-Dickstein, Jascha; Weiss, Eric; Maheswaranathan, Niru; Ganguli, Surya (2015-06-01).
5124:
3223:
2881:
2763:
2736:
2690:
2496:
2321:
2062:
1657:
1625:
1440:
1234:
1212:
1158:
1135:
1068:
also published adversarial neural networks that contest with each other in the form of a
890:
769:
710:
644:
609:
568:
505:
469:
450:
257:
13622:"Are there Deep Reasons Underlying the Pathologies of Today's Deep Learning Algorithms?"
13280:
13210:
12821:
12593:
12311:
12119:
12052:
12037:"Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations"
11995:
11930:
11812:
11785:
11769:
11645:
11474:
11363:
10093:"A Latent Semantic Model with Convolutional-Pooling Structure for Information Retrieval"
9415:
9105:
8567:
8312:
8026:
7343:
7057:
6981:
6422:
6149:
5736:
Diploma thesis. Institut f. Informatik, Technische Univ. Munich. Advisor: J. Schmidhuber
5698:
5471:
5424:
5377:
5287:
5266:
5236:
4537:
4439:
4153:
3976:
3803:
3332:
1165:
reported significant success with deep neural networks in speech processing in the 1998
14936:
14514:
14097:
13960:
13674:
13653:
13417:
13308:
13232:
13139:
13116:
13104:
13085:
13072:
13039:
13020:
12969:
12918:
12861:
12848:
12805:
12786:
12735:
12722:
12687:
12668:
12545:
12430:
12297:
12265:
12231:
12200:
12174:
12138:
12105:
12093:
12069:
12036:
12017:
11960:
11860:
11695:
11596:
11550:
11515:
11496:
11457:
Dong, Xin; Zhou, Yizhao; Wang, Lantian; Peng, Jingfeng; Lou, Yanbo; Fan, Yiqun (2020).
11428:
11383:
11349:
11321:
11295:
11205:
11172:
11153:
11140:
11113:
11082:
10995:
10966:
10911:
10862:
10817:
10759:
10731:
10542:
10413:
10289:
10262:
10243:
10174:
10055:
9962:
9848:
9753:
9688:
9644:
9602:
9495:
9469:
9432:
9399:
9318:
9117:
9091:
9052:
9024:
8909:
8871:
8848:
8800:
8728:
8663:
8623:
8599:
8553:
8507:
8340:
8278:
8179:
8135:
8107:
8041:
7979:
7897:
7858:
7827:
7802:
7782:
7756:
7734:
7711:
7690:
7658:
7637:
7499:
7465:
7399:
7267:
7069:
6821:
6754:
6442:
6377:
6169:
6100:
6014:"Chapter 6: Information Processing in Dynamical Systems: Foundations of Harmony Theory"
5941:
5915:
5882:
5679:
5535:
5094:
4981:
4938:
4897:
4874:
4522:"Neural networks and physical systems with emergent collective computational abilities"
4500:
4343:
4325:
4172:
4137:
4035:
3988:
3931:
3704:
3678:
3636:
3610:
3559:
3406:
3378:
3352:
3291:
3203:
3193:
3170:
Mühlhoff argues that in most commercial end-user applications of Deep Learning such as
3152:
2877:
2525:
2452:
2393:
2344:
2265:
2250:
2093:
2056:
Feature processing by deep models with solid understanding of the underlying mechanisms
1907:
1822:
1621:
1227:
1142:
1127:
939:
905:
882:
867:
818:
773:
699:
675:
497:
307:
13930:
12897:
12880:
12218:
Bernhardt, Melanie; Vishnevskiy, Valery; Rau, Richard; Goksel, Orcun (December 2020).
10063:
Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing
7427:. ICML '09. New York, NY, USA: Association for Computing Machinery. pp. 873–880.
6556:
6526:
6213:
5976:
5502:
4556:
4521:
4018:
Hornik, Kurt (1991). "Approximation Capabilities of Multilayer Feedforward Networks".
3313:
3128:
2380:
variables. The estimated value function was shown to have a natural interpretation as
1368:
The success in image classification was then extended to the more challenging task of
15002:
14990:
14794:
14446:
14317:
14310:
14072:
14044:
14025:
13986:
13964:
13952:
13407:
13300:
13292:
13224:
13144:
13077:
13059:
13012:
13004:
12961:
12953:
12910:
12902:
12853:
12835:
12804:
Buesing, Lars; Bill, Johannes; Nessler, Bernhard; Maass, Wolfgang (3 November 2011).
12778:
12770:
12727:
12709:
12660:
12625:
12620:
12607:
12577:
12537:
12492:
12488:
12455:
12422:
12333:
12325:
12269:
12257:
12249:
12204:
12192:
12143:
12074:
12021:
12009:
11952:
11944:
11864:
11852:
11836:
11817:
11799:
11699:
11586:
11555:
11537:
11500:
11488:
11418:
11403:"Deep Convolutional Neural Networks for Detecting Cellular Changes Due to Malignancy"
11375:
11325:
11313:
11210:
11192:
11145:
11086:
11072:
11000:
10866:
10854:
10593:
10534:
10294:
9829:
9692:
9606:
9528:
9499:
9487:
9437:
9042:
8994:
8899:
8838:
8749:
8720:
8712:
8499:
8332:
8324:
8264:
8139:
8125:
7876:
7618:
7608:
7491:
7483:
7436:
7137:
7073:
6995:
6946:
6938:
6899:
6813:
6788:
6746:
6738:
6579:
6499:
6369:
6306:
6276:
6241:
6233:
6229:
6214:"Predicting the secondary structure of globular proteins using neural network models"
6192:
6161:
6092:
6087:
6030:
5980:
5945:
5933:
5821:
5757:
5631:
5604:
5483:
5436:
5389:
5248:
5199:
5098:
5006:
4985:
4973:
4942:
4930:
4777:
4740:
4679:
4671:
4592:
4561:
4371:
4266:
4177:
4093:
4063:
4031:
3935:
3923:
3864:
3729:
3696:
3628:
3563:
3396:
3344:
3295:
3283:
3238:
2984:
2914:
2873:
2743:
somewhat analogous to the neural networks utilized in deep learning models. Like the
2740:
2727:
2715:
2599:
2595:
2590:
2554:
2066:
1925:
1809:
1773:-regularization) can be applied during training to combat overfitting. Alternatively
1700:
As with ANNs, many issues can arise with naively trained DNNs. Two common issues are
1586:
1146:
1112:
1104:
1012:
971:
851:
741:
586:
572:
532:
445:
85:
13089:
13024:
12973:
12922:
12790:
11964:
11600:
10546:
10382:"Zero-Shot Translation with Google's Multilingual Neural Machine Translation System"
10247:
9648:
9121:
8913:
8852:
8804:
8667:
8511:
8121:
6758:
6461:
5750:"Gradient flow in recurrent nets: the difficulty of learning long-term dependencies"
5683:
5539:
4347:
3708:
2987:
until it tricked the anti-malware while retaining its ability to damage the target.
2852:
Others point out that deep learning should be looked at as a step towards realizing
2845:
how fast? What is it approximating?) Deep learning methods are often looked at as a
1531:
14747:
14737:
14544:
14338:
14288:
14283:
14226:
14214:
13942:
13874:
13713:
13421:
13399:
13284:
13267:
13236:
13214:
13134:
13130:
13126:
13067:
13051:
12996:
12945:
12892:
12865:
12843:
12825:
12762:
12739:
12717:
12699:
12672:
12652:
12615:
12597:
12549:
12529:
12484:
12434:
12414:
12385:
12315:
12241:
12184:
12133:
12123:
12064:
12056:
11999:
11934:
11844:
11807:
11789:
11687:
11578:
11545:
11527:
11478:
11432:
11410:
11387:
11367:
11305:
11229:"DeepMind's protein-folding AI has solved a 50-year-old grand challenge of biology"
11200:
11184:
11157:
11135:
11125:
11064:
11056:
10990:
10980:
10846:
10583:
10573:
10524:
10439:
10284:
10274:
10235:
10178:
10166:
9941:
9903:
9821:
9678:
9636:
9594:
9563:
9520:
9479:
9427:
9419:
9328:
9109:
9056:
9034:
8891:
8830:
8792:
8732:
8704:
8695:
Hochreiter, Sepp; Schmidhuber, Jürgen (1 November 1997). "Long Short-Term Memory".
8655:
8491:
8480:"LSTM Recurrent Networks Learn Simple Context Free and Context Sensitive Languages"
8316:
8117:
7868:
7600:
7529:
7519:"Flexible, High Performance Convolutional Neural Networks for Image Classification"
7518:
7503:
7475:
7428:
7347:
7259:
7061:
6985:
6930:
6891:
6825:
6805:
6730:
6552:
6522:
6495:
6434:
6426:
6381:
6361:
6353:
6268:
6225:
6153:
6104:
6082:
6074:
5972:
5925:
5874:
5813:
5791:
5671:
5656:
5596:
5527:
5475:
5428:
5381:
5240:
5086:
5074:
5056:
5038:
4965:
4922:
4854:
4806:
4773:
4700:
4663:
4617:
4584:
4551:
4541:
4443:
4335:
4298:
4220:
4167:
4157:
4039:
4027:
3992:
3980:
3913:
3856:
3807:
3688:
3620:
3551:
3410:
3388:
3356:
3336:
3275:
3160:
2805:
2759:
2684:
2558:
2546:
2538:
2456:
2373:
2271:
2101:
2076:
1921:
1433:
1358:
1081:
1073:
931:
893:. The rectifier has become the most popular activation function for deep learning.
836:
805:(RNN). RNNs have cycles in their connectivity structure, FNNs don't. In the 1920s,
729:
694:
636:
579:
564:
429:
223:
158:
143:
13312:
12161:
Oktem, Figen S.; Kar, Oğuzhan Fatih; Bezek, Can Deniz; Kamalabadi, Farzad (2021).
11173:"Using recurrent neural network models for early detection of heart failure onset"
9640:
8659:
8344:
7271:
7263:
7252:"Conversational speech transcription using context-dependent deep neural networks"
6680:
6446:
6173:
5886:
5865:(2010). "Formal Theory of Creativity, Fun, and Intrinsic Motivation (1990-2010)".
4795:
3996:
3918:
3901:
3640:
1928:-sequence recognition, which, unlike word-sequence recognition, allows weak phone
608:. Prior to deep learning, machine learning techniques often involved hand-crafted
14860:
14804:
14626:
14268:
14188:
12830:
12449:
11691:
10578:
10561:
10124:"Learning Deep Structured Semantic Models for Web Search using Clickthrough Data"
10019:
9825:
9512:
9374:"Cerebras launches new AI supercomputing processor with 2.6 trillion transistors"
9152:
7604:
7517:
Ciresan, D. C.; Meier, U.; Masci, J.; Gambardella, L.M.; Schmidhuber, J. (2011).
7351:
7158:"Deng receives prestigious IEEE Technical Achievement Award - Microsoft Research"
7131:
7023:
6861:
6687:
6462:"Artificial Neural Networks and their Application to Speech/Sequence Recognition"
6125:
6058:
6054:
5929:
5777:
5749:
5731:
5326:"Shift-invariant pattern recognition neural network and its optical architecture"
5273:
5026:
4838:
4734:
4588:
4260:
4123:
4057:
3837:
3692:
3156:
2957:
2755:
2582:
2570:
2460:
2137:
1834:
1620:
Neural networks have been used on a variety of tasks, including computer vision,
1614:
1477:
1425:
1342:
1334:
1289:
1223:
1100:
1092:
1046:
923:
911:
659:
493:
100:
13403:
12376:
Galkin, F.; Mamoshina, P.; Kochetov, K.; Sidorenko, D.; Zhavoronkov, A. (2020).
11745:"Google DeepMind's materials AI has already discovered 2.2 million new crystals"
11582:
11483:
11458:
11371:
9567:
8895:
8834:
6934:
6809:
3464:"Google's AlphaGo AI wins three-match series against the world's best Go player"
1277:
The deep learning revolution started around CNN- and GPU-based computer vision.
14834:
14799:
14789:
14614:
14372:
14198:
14058:
13262:
13258:
12949:
12245:
12219:
12162:
11848:
11794:
11309:
11171:
Choi, Edward; Schuetz, Andy; Stewart, Walter F.; Sun, Jimeng (13 August 2016).
10279:
10170:
10056:"Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank"
9774:"MNIST handwritten digit database, Yann LeCun, Corinna Cortes and Chris Burges"
9598:
9483:
9028:
8750:"Learning Precise Timing with LSTM Recurrent Networks (PDF Download Available)"
8708:
8104:
2021 International Conference on Computer Communication and Informatics (ICCCI)
7848:
7369:
6895:
6734:
6701:
Offline Handwriting Recognition with Multidimensional Recurrent Neural Networks
5600:
4339:
4162:
3144:
2707:
2479:
over the possible classes of random variable Y, given input X. For example, in
2418:
2408:
2305:
2261:
2232:
2167:
1883:
1853:
1790:
1689:
1679:
1629:
1543:
1459:
1405:
1354:
1338:
1242:
1072:, where one network's gain is the other network's loss. The first network is a
927:
765:
745:
690:
509:
13878:
13556:
12766:
12656:
12533:
12418:
12004:
11979:
11939:
11914:
10850:
10469:
Boitet, Christian; Blanchon, Hervé; Seligman, Mark; Bellynck, Valérie (2010).
9683:
9423:
9113:
6990:
6965:
6681:
An application of recurrent neural networks to discriminative keyword spotting
6365:
6272:
6078:
5878:
5675:
5219:
Rumelhart, David E.; Hinton, Geoffrey E.; Williams, Ronald J. (October 1986).
4859:
4842:
4810:
4621:
4447:
3860:
3812:
3787:
3392:
3279:
1664:
are exponentially easier to approximate with DNNs than with shallow networks.
15039:
14779:
14759:
14676:
14062:
13990:
13956:
13947:
13296:
13063:
13008:
12957:
12906:
12839:
12774:
12713:
12704:
12664:
12611:
12329:
12253:
12196:
12188:
12094:"Solving high-dimensional partial differential equations using deep learning"
12013:
11948:
11803:
11541:
11492:
11196:
9631:
Deng, L.; Platt, J. (2014). "Ensemble Deep Learning for Speech Recognition".
8868:
2013 IEEE International Conference on Acoustics, Speech and Signal Processing
8827:
2013 IEEE International Conference on Acoustics, Speech and Signal Processing
8780:
8716:
8479:
8328:
7487:
7065:
6999:
6942:
6742:
6373:
6337:
Waibel, A.; Hanazawa, T.; Hinton, G.; Shikano, K.; Lang, K. J. (March 1989).
6280:
6237:
6062:
6020:
5984:
5608:
5252:
4675:
4302:
3927:
3287:
3233:
3112:
2796:
video games using only pixels as data input. In 2015 they demonstrated their
2422:
1858:
1782:
1743:
1678:, in which data can flow in any direction, are used for applications such as
1564:
1473:
1428:(2015) eclipsed GANs in generative modeling since then, with systems such as
1258:
1131:
1069:
822:
761:
725:
13261:; Lillicrap, Timothy; Leach, Madeleine; Kavukcuoglu, Koray; Graepel, Thore;
12602:
12320:
12285:
12128:
12060:
11532:
11060:
11050:
10641:"Multi-task Neural Networks for QSAR Predictions | Data Science Association"
9847:
Chaochao Lu; Xiaoou Tang (2014). "Surpassing Human Level Face Recognition".
9333:
9306:
9038:
7432:
7133:
Automatic Speech Recognition: A Deep Learning Approach (Publisher: Springer)
6707:, Neural Information Processing Systems (NIPS) Foundation, 2009, pp. 545–552
6430:
6298:
6157:
5959:
Ackley, David H.; Hinton, Geoffrey E.; Sejnowski, Terrence J. (1985-01-01).
5561:
5500:
3848:
14865:
14696:
14111:
13304:
13228:
13148:
13081:
13055:
13016:
12965:
12936:
Olshausen, B; Field, D (1 August 2004). "Sparse coding of sensory inputs".
12914:
12857:
12782:
12731:
12541:
12496:
12426:
12337:
12261:
12147:
12078:
11856:
11821:
11653:. Computer Vision and Pattern Recognition (CVPR), 2014 IEEE Conference on.
11559:
11379:
11317:
11214:
11188:
11149:
11004:
10858:
10597:
10538:
10344:"Found in translation: More accurate, fluent sentences in Google Translate"
10298:
9833:
9491:
9441:
8818:
8503:
8336:
8235:
7622:
7495:
7425:
Proceedings of the 26th Annual International Conference on Machine Learning
7395:
7041:
6950:
6903:
6817:
6784:
6750:
6133:
5937:
5440:
5409:"Image processing of human corneal endothelium based on a learning network"
5393:
5220:
4683:
4651:
4579:
Nakano, Kaoru (1971). "Learning Process in a Model of Associative Memory".
4546:
4181:
3826:
3700:
3632:
3348:
3164:
3132:
3123:) is regularly deployed for this purpose, but also implicit forms of human
3116:
2906:
2891:
2242:
2191:
Visual art processing of Jimmy Wales in France, with the style of Munch's "
2160:
2105:
1879:
1712:
1602:
1575:
1485:
1231:
1193:
970:
on mail. Training required 3 days. In 1990, Wei Zhang implemented a CNN on
901:
806:
733:
655:
441:
297:
13621:
12629:
11414:
9257:
8724:
8415:
7420:
6657:
Proceedings of the International Conference on Machine Learning, ICML 2006
6612:
Graves, Alex; Eck, Douglas; Beringer, Nicole; Schmidhuber, Jürgen (2003).
6245:
6165:
6096:
5817:
5795:
5562:"Attractor dynamics and parallelism in a connectionist sequential machine"
5487:
4977:
4934:
4565:
3624:
2714:. The aging clock was planned to be released for public use in 2021 by an
14961:
14732:
14641:
14636:
14258:
14236:
11254:"DeepMind solves 50-year-old 'grand challenge' with protein folding A.I."
11130:
10263:"Precision information extraction for rare disease epidemiology at scale"
10195:
Gao, Jianfeng; He, Xiaodong; Yih, Scott Wen-tau; Deng, Li (1 June 2014).
8024:
7872:
7479:
6299:"A real-time recurrent error propagation network word recognition system"
6129:
6050:
5432:
5385:
5305:
5154:
3749:
Bengio, Yoshua; Lamblin, Pascal; Popovici, Dan; Larochelle, Hugo (2007).
3669:
Schmidhuber, J. (2015). "Deep Learning in Neural Networks: An Overview".
2857:
2698:
and predicts people with certain conditions older than healthy controls:
2439:, according to the sequence of the amino acids that make it up. In 2020,
2414:
2340:
2336:
1701:
1362:
1308:
1173:
1096:
955:
935:
830:
814:
810:
536:
521:
513:
326:
311:
13288:
11068:
9946:
9929:
9908:
9891:
9722:"How Skype Used AI to Build Its Amazing New Language Translator | WIRED"
9664:"Phone Recognition with Hierarchical Convolutional Deep Maxout Networks"
9514:
8320:
7597:
Medical Image Computing and Computer-Assisted Intervention – MICCAI 2013
6438:
6261:
International Journal of Pattern Recognition and Artificial Intelligence
5330:
Proceedings of Annual Conference of the Japan Society of Applied Physics
3340:
2634:
926:
had a continuous precursor of backpropagation in 1960 in the context of
900:(CNNs) with convolutional layers and downsampling layers began with the
14855:
14814:
14809:
14722:
14631:
14539:
14451:
14431:
13717:
12224:
IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control
12035:
Raissi, Maziar; Yazdani, Alireza; Karniadakis, George Em (2020-02-28).
11407:
2017 IEEE International Conference on Computer Vision Workshops (ICCVW)
10053:
10018:
Socher, Richard; Bauer, John; Manning, Christopher; Ng, Andrew (2013).
9524:
7599:. Lecture Notes in Computer Science. Vol. 7908. pp. 411–418.
7016:
7012:
5090:
4969:
4926:
4224:
3984:
3758:. Advances in neural information processing systems. pp. 153–160.
3555:
3039: in this section. Unsourced material may be challenged and removed.
2991:
2990:
In 2016, another group demonstrated that certain sounds could make the
2849:, with most confirmations done empirically, rather than theoretically.
2748:
2550:
2399:
content-based approach and enhances recommendations in multiple tasks.
2332:
of environmental chemicals in nutrients, household products and drugs.
2325:
2309:
2192:
2109:
1875:
1518:
1481:
1444:
1207:, Santiago Fernández, Faustino Gomez, and Schmidhuber combined it with
1162:
979:
959:
942:
et al. popularised backpropagation but did not cite the original work.
934:'s master thesis (1970). G.M. Ostrovski et al. republished it in 1971.
915:
525:
11956:
10588:
8495:
7855:
2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
7371:
High performance convolutional neural networks for document processing
41:
Representing images on multiple layers of abstraction in deep learning
14850:
14819:
14717:
14561:
14524:
14461:
14415:
14410:
14395:
13498:"DARPA is funding projects that will try to open up AI's black boxes"
13254:
12282:
11978:
Mao, Zhiping; Jagtap, Ameya D.; Karniadakis, George Em (2020-03-01).
10611:
10239:
10197:"Learning Continuous Phrase Representations for Translation Modeling"
9773:
9304:
8796:
7822:
He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian (10 Dec 2015).
7330:
Oh, K.-S.; Jung, K. (2004). "GPU implementation of neural networks".
6357:
5810:
9th International Conference on Artificial Neural Networks: ICANN '99
5531:
5479:
5244:
4667:
3124:
2972:
2965:
2869:
2865:
2846:
2801:
2744:
2542:
2440:
1890:
1387:
1383:(ResNet) in Dec 2015. ResNet behaves like an open-gated Highway Net.
1315:
1285:
1057:
connections to solve the vanishing gradient problem. This led to the
821:
made this architecture adaptive. His learning RNN was republished by
681:
The classic universal approximation theorem concerns the capacity of
582:
361:
125:
13219:
13194:
13000:
12562:
S. Blakeslee, "In brain's early growth, timetable may be critical",
10529:
10512:
10380:
Schuster, Mike; Johnson, Melvin; Thorat, Nikhil (22 November 2016).
8034:
Proceedings of the 32nd International Conference on Machine Learning
4762:"Heuristic self-organization in problems of engineering cybernetics"
3436:
NIPS 2012: Neural Information Processing Systems, Lake Tahoe, Nevada
3014:
2808:
uses a neural network to translate between more than 100 languages.
2721:
1211:(CTC) in stacks of LSTMs. In 2009, it became the first RNN to win a
14752:
14584:
13904:
13767:"Hackers Have Already Started to Weaponize Artificial Intelligence"
13557:
Alexander Mordvintsev; Christopher Olah; Mike Tyka (17 June 2015).
12302:
12236:
12179:
12110:
11354:
11300:
10916:
10822:
10736:
10418:
9582:
9474:
9323:
9096:
8929:"Improving DNNs for LVCSR using rectified linear units and dropout"
8628:
8604:
8112:
8046:
7984:
7902:
7863:
7832:
7807:
7594:
7560:
7404:
7298:"Recent Advances in Deep Learning for Speech Research at Microsoft"
5920:
5042:
4902:
4612:
Nakano, Kaoru (1972). "Associatron-A Model of Associative Memory".
4505:
4330:
3171:
3148:
2778:
2695:
2492:
2329:
2317:
2237:
1868:
1846:
1688:(CNNs) are used in computer vision. CNNs also have been applied to
1539:
1514:
1421:
1417:
1181:
850:, a method to train arbitrarily deep neural networks, published by
737:
198:
120:
14066:
13679:
13658:
13396:
2008 7th IEEE International Conference on Development and Learning
13121:
12378:"DeepMAge: A Methylation Aging Clock Developed with Deep Learning"
10985:
10755:"Toronto startup has a faster way to discover effective medicines"
10435:"An Infusion of AI Makes Google Translate More Powerful Than Ever"
9967:
9853:
9758:
9167:"Deep Neural Networks for Acoustic Modeling in Speech Recognition"
8876:
8865:
8817:
8558:
8184:
7787:
7779:
Very Deep Convolutional Networks for Large-Scale Image Recognition
7761:
7739:
7716:
7695:
7663:
7642:
7470:
7455:
7421:"Large-scale deep unsupervised learning using graphics processors"
5566:
Proceedings of the Annual Meeting of the Cognitive Science Society
3683:
3615:
3383:
2979:
defense industry. ANNs have been trained to defeat ANN-based anti-
1808:
Since the 2010s, advances in both machine learning algorithms and
1126:
Both shallow and deep learning (e.g., recurrent nets) of ANNs for
14875:
14712:
14666:
14589:
14489:
14484:
14436:
13694:
Zhu, S.C.; Mumford, D. (2006). "A stochastic grammar of images".
13365:"A.I. Researchers Leave Elon Musk Lab to Begin Robotics Start-Up"
12578:"A more biologically plausible learning rule for neural networks"
12451:
Rethinking Innateness: A Connectionist Perspective on Development
12405:
Utgoff, P. E.; Stracuzzi, D. J. (2002). "Many-layered learning".
11715:"Deep learning: the next frontier for money laundering detection"
11114:"Sleep Quality Prediction From Wearable Data Using Deep Learning"
6917:
Hinton, Geoffrey E.; Osindero, Simon; Teh, Yee-Whye (July 2006).
6614:"Biologically Plausible Speech Recognition with LSTM Neural Nets"
6512:
5776:
5501:
LeCun, Yann; Léon Bottou; Yoshua Bengio; Patrick Haffner (1998).
5346:, "Backpropagation Applied to Handwritten Zip Code Recognition",
3494:"Study urges caution when comparing neural networks to the brain"
3429:"ImageNet Classification with Deep Convolutional Neural Networks"
3136:
2980:
2976:
2797:
2711:
2488:
2121:
2092:
All major commercial speech recognition systems (e.g., Microsoft
1917:
1682:. Long short-term memory is particularly effective for this use.
1598:
1372:(captions) for images, often as a combination of CNNs and LSTMs.
1330:
1323:
1040:
366:
12881:"Linear summation of excitatory inputs by CA1 pyramidal neurons"
12390:
12377:
12375:
11400:
9889:
9307:"In-Datacenter Performance Analysis of a Tensor Processing Unit"
8597:
8027:"Deep Unsupervised Learning using Nonequilibrium Thermodynamics"
7226:
NIPS Workshop on Deep Learning and Unsupervised Feature Learning
6679:
Santiago Fernandez, Alex Graves, and Jürgen Schmidhuber (2007).
2213:
generating striking imagery based on random visual input fields.
1007:
proposed a hierarchy of RNNs pre-trained one level at a time by
37:
14890:
14870:
14742:
14534:
10559:
10513:"Trial watch: Phase II and phase III attrition rates 2011-2012"
9587:
IEEE/ACM Transactions on Audio, Speech, and Language Processing
9459:
8416:"The power of deeper networks for expressing natural functions"
6647:
3748:
3375:
2012 IEEE Conference on Computer Vision and Pattern Recognition
2947:
2507:
and outperforms other methods in case of large alphabet sizes.
2484:
1929:
1830:
1814:
1590:
1429:
1413:
1395:
1350:
670:
Deep neural networks are generally interpreted in terms of the
11767:
10468:
8384:
Szegedy, Christian; Toshev, Alexander; Erhan, Dumitru (2013).
7916:
7847:
He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian (2016).
5265:
Rumelhart, David E., Geoffrey E. Hinton, and R. J. Williams. "
5077:(1976). "Taylor expansion of the accumulated rounding error".
4117:
The Expressive Power of Neural Networks: A View from the Width
3603:
IEEE Transactions on Pattern Analysis and Machine Intelligence
2451:
Deep neural networks can be used to estimate the entropy of a
752:. The probabilistic interpretation led to the introduction of
14691:
14671:
14661:
14656:
14651:
14646:
14609:
14441:
13529:"Is "Deep Learning" a Revolution in Artificial Intelligence?"
13528:
12217:
11913:
Raissi, M.; Perdikaris, P.; Karniadakis, G. E. (2019-02-01).
11018:
Elkahky, Ali Mamdouh; Song, Yang; He, Xiaodong (1 May 2015).
10930:
10225:
9197:"GPUs Continue to Dominate the AI Accelerator Market for Now"
8995:"A Practical Guide to Training Restricted Boltzmann Machines"
8621:
6346:
IEEE Transactions on Acoustics, Speech, and Signal Processing
4426:
Brush, Stephen G. (1967). "History of the Lenz-Ising Model".
2793:
2231:
Other key techniques in this field are negative sampling and
2117:
1594:
1455:
1281:
optimizations were developed specifically for deep learning.
1254:
1176:
features in the late 1990s, showing its superiority over the
1154:
975:
963:
598:
590:
13979:"Facebook Can Now Find Your Face, Even When It's Not Tagged"
13672:
13393:
10815:
9801:
9560:
Cambridge University Engineering Department Technical Report
8258:
7516:
6611:
5159:"Applications of advances in nonlinear sensitivity analysis"
5063:(Masters) (in Finnish). University of Helsinki. p. 6–7.
4499:(2022). "Annotated History of Modern AI and Deep Learning".
3427:
Krizhevsky, Alex; Sutskever, Ilya; Hinton, Geoffrey (2012).
2435:
Deep neural networks have shown unparalleled performance in
2320:. Research has explored use of deep learning to predict the
930:. The modern form of backpropagation was first published in
547:
Most modern deep learning models are based on multi-layered
492:. These architectures have been applied to fields including
14681:
12576:
Mazzoni, P.; Andersen, R. A.; Jordan, M. I. (15 May 1991).
12286:"Learning skillful medium-range global weather forecasting"
11912:
10159:
IEEE Transactions on Audio, Speech, and Language Processing
9400:"Logic-in-memory based on an atomically thin semiconductor"
9350:"Cerebras Hits the Accelerator for Deep Learning Workloads"
9025:"Scaling deep learning on GPU and knights landing clusters"
7419:
Raina, Rajat; Madhavan, Anand; Ng, Andrew Y. (2009-06-14).
6336:
2925:
and is a basic goal of both human language acquisition and
2726:
Deep learning is closely related to a class of theories of
2113:
2097:
1467:
1269:
1123:, an early application of deep learning to bioinformatics.
714:
13164:"Facebook's 'Deep Learning' Guru Reveals the Future of AI"
13040:"An emergentist perspective on the origin of number sense"
7731:
7185:
7183:
2971:
ANNs can however be further trained to detect attempts at
1585:
An ANN is based on a collection of connected units called
1521:, which are correlated with "nodes" that represent visual
14057:
13742:"Deep Learning of Recursive Structure: Grammar Induction"
11884:"Army researchers develop new algorithms to train robots"
11837:"Google AI and robots join forces to build new materials"
11288:
IEEE Transactions on Neural Networks and Learning Systems
10882:"A Molecule Designed By AI Exhibits 'Druglike' Qualities"
9750:
9081:
8781:"Gradient-based learning applied to document recognition"
8446:"Is Artificial Intelligence Finally Coming into Its Own?"
8054:
7526:
International Joint Conference on Artificial Intelligence
6258:
5503:"Gradient-based learning applied to document recognition"
4115:
Lu, Z., Pu, H., Wang, F., Hu, Z., & Wang, L. (2017).
4081:
4079:
3956:"Approximations by superpositions of sigmoidal functions"
3849:
Aizenberg, I.N.; Aizenberg, N.N.; Vandewalle, J. (2000).
3426:
2619:
Physics informed neural networks have been used to solve
2569:
Deep learning is being successfully applied to financial
2257:(token classification), text classification, and others.
1288:
reported a 100M deep belief network trained on 30 Nvidia
829:
were published by Kaoru Nakano in 1971. Already in 1948,
685:
with a single hidden layer of finite size to approximate
12034:
11615:"Colorizing and Restoring Old Images with Deep Learning"
11338:
10375:
10373:
10371:
10369:
10260:
10153:
Mesnil, G.; Dauphin, Y.; Yao, K.; Bengio, Y.; Deng, L.;
9868:
Nvidia Demos a Car Computer Trained with "Deep Learning"
9023:
You, Yang; Buluç, Aydın; Demmel, James (November 2017).
8825:(2013). "Deep convolutional neural networks for LVCSR".
8153:
Sak, Hasim; Senior, Andrew; Beaufays, Francoise (2014).
7709:
7368:
Chellapilla, Kumar; Puri, Sidd; Simard, Patrice (2006),
6542:
2503:
holds. It is shown that this method provides a strongly
2296:
post-mortem matching and determination of subject sex.
2287:
English as an intermediate between most language pairs.
2166:
A common evaluation set for image classification is the
2155:
Richard Green explains how deep learning is used with a
1454:(ASR). Results on commonly used evaluation sets such as
464:
Some common deep learning network architectures include
13249:
12803:
12160:
11980:"Physics-informed neural networks for high-speed flows"
11513:
11177:
Journal of the American Medical Informatics Association
8297:
7977:
7754:
7180:
7087:
7085:
7083:
6408:
5218:
3372:
3127:
that are often not recognized as such. The philosopher
2642:
is a numerical method that combines deep learning with
10152:
9846:
9671:
EURASIP Journal on Audio, Speech, and Music Processing
9583:"Convolutional Neural Networks for Speech Recognition"
6339:"Phoneme recognition using time-delay neural networks"
6019:. In Rumelhart, David E.; McLelland, James L. (eds.).
5807:
5267:
Learning Internal Representations by Error Propagation
4895:
4360:
4076:
3600:
3312:
LeCun, Yann; Bengio, Yoshua; Hinton, Geoffrey (2015).
2335:
AtomNet is a deep learning system for structure-based
2050:
Scale-up/out and accelerated DNN training and decoding
635:
Deep learning architectures can be constructed with a
13436:"Talk to the Algorithms: AI Becomes a Faster Learner"
13103:
Güçlü, Umut; van Gerven, Marcel A. J. (8 July 2015).
12752:
12575:
11984:
Computer Methods in Applied Mechanics and Engineering
11170:
10426:
10379:
10366:
9980:
9978:
9927:
8694:
8000:"Prepare, Don't Panic: Synthetic Media and Deepfakes"
7367:
5958:
5700:
Habilitation thesis: System modeling and optimization
5221:"Learning representations by back-propagating errors"
4051:
4049:
2640:
Deep backward stochastic differential equation method
2635:
Deep backward stochastic differential equation method
1752:
1721:
13038:
Zorzi, Marco; Testolin, Alberto (19 February 2018).
11977:
11278:
10934:
Advances in Neural Information Processing Systems 26
10729:
10017:
9890:
G. W. Smith; Frederic Fol Leymarie (10 April 2017).
8543:"Sequence to Sequence Learning with Neural Networks"
8420:
International Conference on Learning Representations
8075:
7564:
Advances in Neural Information Processing Systems 25
7080:
7035:
7033:
7031:
6673:
6485:
6049:
5311:
Phoneme Recognition Using Time-Delay Neural Networks
5289:
Phoneme Recognition Using Time-Delay Neural Networks
4291:
IEEE Transactions on Systems Science and Cybernetics
2730:(specifically, neocortical development) proposed by
2573:, tax evasion detection, and anti-money laundering.
2561:, which trains on the image that needs restoration.
654:
was introduced to the machine learning community by
13651:
11048:
9397:
8101:
7971:
7363:
7361:
7091:
6124:
5029:(1960). "Gradient theory of optimal flight paths".
4877:(1967). "A theory of adaptive pattern classifier".
4126:. Neural Information Processing Systems, 6231-6239.
2792:developed a system capable of learning how to play
2372:has been used to approximate the value of possible
2358:
2128:speech products, etc.) are based on deep learning.
1867:In 2021, J. Feldmann et al. proposed an integrated
1080:over output patterns. The second network learns by
12353:"GraphCast: A breakthrough in Weather Forecasting"
11572:
9975:
8540:
8383:
7588:
5867:IEEE Transactions on Autonomous Mental Development
5650:
5648:
4803:IEEE Transactions on Systems, Man, and Cybernetics
4614:IEEE Transactions on Systems, Man, and Cybernetics
4046:
3139:for image recognition or click-tracking on Google
2921:that operates on concepts in terms of grammatical
2446:
1765:
1734:
846:The first working deep learning algorithm was the
578:Fundamentally, deep learning refers to a class of
13559:"Inceptionism: Going Deeper into Neural Networks"
11677:
10964:
10335:
9954:
9228:"AI is changing the entire nature of computation"
8390:Advances in Neural Information Processing Systems
8152:
7953:"GAN 2.0: NVIDIA's Hyperrealistic Face Generator"
7945:
7777:Simonyan, Karen; Zisserman, Andrew (2015-04-10),
7776:
7028:
6916:
6783:
6212:Qian, Ning; Sejnowski, Terrence J. (1988-08-20).
4695:
4693:
3311:
2722:Relation to human cognitive and brain development
2614:
2516:specialists to improve the diagnosis efficiency.
2495:size of Y. NJEE uses continuously differentiable
1184:, later produced excellent larger-scale results.
1149:in the late 1990s. Funded by the US government's
908:in 1979, though not trained by backpropagation.
563:or latent variables organized layer-wise in deep
15037:
13905:"Whose intelligence is artificial intelligence?"
13797:"How hackers can force AI to make dumb mistakes"
13195:"Google AI algorithm masters ancient game of Go"
11647:Shrinkage Fields for Effective Image Restoration
11456:
10787:"Startup Harnesses Supercomputers to Seek Cures"
10510:
10410:
10404:
9516:TIMIT Acoustic-Phonetic Continuous Speech Corpus
8964:"Data Augmentation - deeplearning.ai | Coursera"
7895:
7358:
7092:Deng, L.; Hinton, G.; Kingsbury, B. (May 2013).
7039:
7006:
6919:"A Fast Learning Algorithm for Deep Belief Nets"
6789:"A Fast Learning Algorithm for Deep Belief Nets"
5839:
5837:
4999:Leibniz, Gottfried Wilhelm Freiherr von (1920).
4284:
4282:
3498:MIT News | Massachusetts Institute of Technology
3422:
3420:
3262:Schulz, Hannes; Behnke, Sven (1 November 2012).
3143:), (3) exploitation of social motivations (e.g.
2491:layer with number of nodes that is equal to the
2299:
2204:identifying the style period of a given painting
1311:CNNs on GPU improved performance significantly.
13102:
12582:Proceedings of the National Academy of Sciences
12509:
12404:
12098:Proceedings of the National Academy of Sciences
11770:"Scaling deep learning for materials discovery"
10669:"Toxicology in the 21st century Data Challenge"
10504:
8645:"Recurrent neural network based language model"
8617:
8615:
8593:
8591:
8477:
8078:"Google voice search: faster and more accurate"
7418:
7394:
5898:
5896:
5752:. In Kolen, John F.; Kremer, Stefan C. (eds.).
5748:Hochreiter, S.; et al. (15 January 2001).
5724:Untersuchungen zu dynamischen neuronalen Netzen
5690:
5645:
4526:Proceedings of the National Academy of Sciences
2537:Deep learning has been successfully applied to
1901:
996:(1990), which applied RNN to study problems in
13467:"In defense of skepticism about deep learning"
12935:
12474:
11111:
10940:. Curran Associates, Inc. pp. 2643–2651.
10705:"NCATS Announces Tox21 Data Challenge Winners"
10680:"NCATS Announces Tox21 Data Challenge Winners"
10228:International Journal of Communication Systems
10146:
10121:
9155:. Neural Processing Letters 22.1 (2005): 1-16.
9022:
8413:
7857:. Las Vegas, NV, USA: IEEE. pp. 770–778.
7656:
7570:. Curran Associates, Inc. pp. 2843–2851.
7295:
7125:
7123:
7121:
6506:
5997:: CS1 maint: DOI inactive as of August 2024 (
5622:
5620:
5618:
4732:
4690:
4635:Turing, Alan (1948). "Intelligent Machinery".
3781:
3779:
3307:
3305:
3219:List of datasets for machine-learning research
2804:well enough to beat a professional Go player.
2217:
2037:Hierarchical Convolutional Deep Maxout Network
1470:. but are more successful in computer vision.
772:and popularized in surveys such as the one by
14127:
14020:Bishop, Christopher M.; Bishop, Hugh (2024).
13836:"AI Is Easy to Fool—Why That Needs to Change"
13037:
12986:
12685:
11279:Shalev, Y.; Painsky, A.; Ben-Gal, I. (2022).
11017:
10553:
10090:
10065:. Association for Computational Linguistics.
7846:
7821:
7800:
6607:
6605:
6211:
5961:"A learning algorithm for boltzmann machines"
5851:. MIT Press/Bradford Books. pp. 222–227.
5834:
5754:A Field Guide to Dynamical Recurrent Networks
5259:
5105:
4279:
4262:Machine Learning: A Probabilistic Perspective
3417:
2425:annotations and gene-function relationships.
1711:methods such as Ivakhnenko's unit pruning or
406:
14141:
14019:
13615:
13613:
13611:
13489:
12642:
10194:
10020:"Parsing With Compositional Vector Grammars"
9960:
9455:
9453:
9451:
8612:
8588:
8536:
8534:
8532:
8478:Gers, Felix A.; Schmidhuber, Jürgen (2001).
8473:
8471:
8360:A Guide to Deep Learning and Neural Networks
8283:: CS1 maint: multiple names: authors list (
7850:Deep Residual Learning for Image Recognition
7840:
7824:Deep Residual Learning for Image Recognition
7325:
7323:
6880:"Learning multiple layers of representation"
6719:"Learning multiple layers of representation"
6287:
5893:
5299:
4837:
4513:
4491:
4489:
4487:
4485:
4483:
4481:
4479:
4459:
4457:
4403:
4401:
4315:
3964:Mathematics of Control, Signals, and Systems
3596:
3594:
3592:
3590:
3588:
3520:
3518:
3516:
3514:
3261:
2564:
2524:Finding the appropriate mobile audience for
1299:by Dan Ciresan, Ueli Meier, Jonathan Masci,
689:. In 1989, the first proof was published by
12091:
10836:
10749:
10747:
9928:Blaise Agüera y Arcas (29 September 2017).
9624:
9580:
9145:Continuous CMAC-QRLS and its systolic array
8821:; Mohamed, Abdel-Rahman; Kingsbury, Brian;
8541:Sutskever, L.; Vinyals, O.; Le, Q. (2014).
8386:"Deep neural networks for object detection"
8071:
8069:
7118:
6538:
6536:
6191:. Cambridge, Massachusetts: The MIT Press.
5902:
5861:
5843:
5718:
5716:
5714:
5696:
5654:
5626:
5615:
5123:
5119:
5117:
4992:
4831:
4789:
4787:
4495:
4318:Applied and Computational Harmonic Analysis
4213:Foundations and Trends in Signal Processing
4199:
4197:
4195:
4193:
4191:
4059:Neural Networks: A Comprehensive Foundation
3820:
3776:
3752:Greedy layer-wise training of deep networks
3668:
3302:
1264:
878:the currently dominant training technique.
14134:
14120:
13693:
13586:"Yes, androids do dream of electric sheep"
12556:
12167:IEEE Transactions on Computational Imaging
10672:
10604:
10464:
10462:
9766:
7748:
7725:
7703:
7213:
6854:Learning multiple layers of representation
6846:
6641:
6602:
6453:
5747:
5279:
5187:
5147:
5073:
5055:
5019:
4793:
4759:
4711:
4699:
4649:
4628:
4605:
4572:
4089:Fundamentals of Artificial Neural Networks
3742:
3664:
3662:
3660:
3658:
3656:
3654:
3652:
3650:
3534:Foundations and Trends in Machine Learning
2290:
2059:Adaptation of DNNs and related deep models
2013:Convolutional DNN w. Heterogeneous Pooling
413:
399:
13946:
13707:
13678:
13657:
13608:
13583:
13218:
13155:
13138:
13120:
13071:
12896:
12878:
12847:
12829:
12721:
12703:
12619:
12601:
12523:
12389:
12350:
12319:
12301:
12235:
12178:
12137:
12127:
12109:
12068:
12003:
11938:
11811:
11793:
11549:
11531:
11482:
11353:
11299:
11274:
11272:
11204:
11139:
11129:
10994:
10984:
10915:
10909:
10821:
10735:
10587:
10577:
10528:
10417:
10288:
10278:
10219:
9984:
9966:
9945:
9907:
9852:
9840:
9815:
9757:
9682:
9630:
9574:
9473:
9448:
9431:
9391:
9332:
9322:
9095:
8986:
8920:
8885:
8875:
8636:
8627:
8603:
8557:
8529:
8468:
8183:
8171:
8111:
8045:
7983:
7901:
7862:
7831:
7806:
7786:
7760:
7738:
7715:
7694:
7662:
7641:
7469:
7403:
7320:
7249:
6989:
6664:
6402:
6186:
6086:
6011:
5919:
5521:
5336:
4955:
4916:
4901:
4867:
4858:
4555:
4545:
4504:
4476:
4454:
4398:
4329:
4288:
4254:
4252:
4250:
4248:
4206:"Deep Learning: Methods and Applications"
4171:
4161:
4111:
4109:
3917:
3852:Multi-Valued and Universal Binary Neurons
3811:
3682:
3614:
3585:
3545:
3511:
3382:
3099:Learn how and when to remove this message
3004:
2868:(...) have no obvious ways of performing
2644:Backward stochastic differential equation
2510:
2387:
728:interpretation derives from the field of
13928:
13619:
12686:Testolin, Alberto; Zorzi, Marco (2016).
12468:
12085:
11112:Sathyanarayana, Aarti (1 January 2016).
10744:
10341:
10190:
10188:
9930:"Art in the Age of Machine Intelligence"
9923:
9921:
9919:
9892:"The Machine as Artist: An Introduction"
9885:
9883:
9881:
9719:
9554:
9164:
8772:
8351:
8192:
8146:
8066:
7910:
7671:
7388:
6777:
6690:. Proceedings of ICANN (2), pp. 220–229.
6651:; Fernández, Santiago; Gomez, Faustino;
6533:
6479:
6293:
5801:
5741:
5711:
5114:
5067:
5049:
4949:
4910:
4784:
4726:
4581:Pattern Recognition and Machine Learning
4519:
4364:Pattern Recognition and Machine Learning
4309:
4188:
3209:List of artificial intelligence projects
2825:
2280:Google Neural Machine Translation (GNMT)
2186:
2182:
2141:
1798:cerebellar model articulation controller
1692:for automatic speech recognition (ASR).
1268:
732:. It features inference, as well as the
709:The universal approximation theorem for
36:
14022:Deep learning: foundations and concepts
13924:
13922:
13920:
13918:
12692:Frontiers in Computational Neuroscience
12503:
12398:
11878:
11876:
11874:
11643:
10459:
9985:Socher, Richard; Manning, Christopher.
8744:
8742:
8642:
7815:
7677:
7629:
6699:Graves, Alex; and Schmidhuber, Jürgen;
5855:
5494:
4998:
4733:Ivakhnenko, A. G.; Lapa, V. G. (1967).
4419:
4135:
4085:
4013:
4011:
4009:
3953:
3899:
3721:
3647:
3184:Applications of artificial intelligence
2968:and caused an ANN to misclassify them.
2783:automatically tagging uploaded pictures
2660:
2463:. Practically, the DNN is trained as a
2075:and how to design them to best exploit
1643:
1039:in an RNN unfolded in time. The "P" in
14:
15038:
14038:
13902:
13864:
13830:
13828:
13826:
13824:
13822:
13526:
13495:
13464:
13192:
12441:
12351:Sivakumar, Ramakrishnan (2023-11-27).
11834:
11269:
10663:
10047:
10027:Proceedings of the ACL 2013 Conference
9744:
9655:
9506:
9347:
9311:ACM SIGARCH Computer Architecture News
8992:
8236:"2018 ACM A.M. Turing Award Laureates"
8198:
7794:
7635:
7329:
7219:
7129:
6963:
6877:
6717:Hinton, Geoffrey E. (1 October 2007).
6716:
6459:
5559:
5353:
5317:
5285:
5193:
5153:
5025:
4796:"Polynomial theory of complex systems"
4736:Cybernetics and Forecasting Techniques
4717:
4634:
4611:
4578:
4258:
4245:
4203:
4106:
4055:
4017:
3949:
3947:
3945:
3785:
3524:
3151:to obtain labeled facial images), (4)
2785:with the names of the people in them.
2773:
1569:are computing systems inspired by the
1161:. The speaker recognition team led by
14115:
13867:"The scientist who spots fake videos"
12879:Cash, S.; Yuste, R. (February 1999).
12447:
11742:
11712:
11251:
10797:from the original on 24 December 2015
10185:
10072:from the original on 28 December 2016
9916:
9878:
9720:McMillan, Robert (17 December 2014).
9581:Abdel-Hamid, O.; et al. (2014).
9137:
9128:
8778:
8414:Rolnick, David; Tegmark, Max (2018).
8259:Ferrie, C., & Kaiser, S. (2019).
7933:from the original on 22 November 2019
7534:10.5591/978-1-57735-516-8/ijcai11-210
5582:
5453:
5406:
5359:
5323:
4873:
4847:The Annals of Mathematical Statistics
4463:
4425:
2669:
2587:Lawrence Berkeley National Laboratory
2519:
2195:" applied using neural style transfer
1640:recognizing faces, or playing "Go").
1209:connectionist temporal classification
14972:Generative adversarial network (GAN)
13929:Mühlhoff, Rainer (6 November 2019).
13915:
13846:from the original on 11 October 2017
13807:from the original on 11 October 2019
13777:from the original on 11 October 2019
13508:from the original on 4 November 2019
13477:from the original on 12 October 2018
13362:
13161:
12092:Han, J.; Jentzen, A.; E, W. (2018).
11871:
11625:from the original on 11 October 2019
11030:from the original on 25 January 2018
10972:Journal of Medical Internet Research
10767:from the original on 20 October 2015
10697:
10612:"Merck Molecular Activity Challenge"
10447:from the original on 8 November 2020
10432:
10342:Turovsky, Barak (15 November 2016).
10207:from the original on 27 October 2017
10134:from the original on 27 October 2017
10103:from the original on 27 October 2017
9661:
9177:from the original on 1 February 2016
8974:from the original on 1 December 2017
8926:
8739:
8484:IEEE Transactions on Neural Networks
8177:
8006:from the original on 2 December 2020
7650:
7308:from the original on 12 October 2017
6878:Hinton, Geoffrey E. (October 2007).
6765:from the original on 11 October 2013
6043:
5812:. Vol. 1999. pp. 850–855.
5131:. IDSIA, Switzerland. Archived from
4354:
4006:
3900:Fradkov, Alexander L. (2020-01-01).
3842:
3527:"Learning Deep Architectures for AI"
3368:
3366:
3189:Comparison of deep learning software
3037:adding citations to reliable sources
3008:
2576:
2532:
2131:
1965:Hidden Trajectory (Generative) Model
1787:Sweeping through the parameter space
793:of artificial neural network (ANN):
13997:from the original on 10 August 2019
13819:
13446:from the original on 28 August 2018
12564:The New York Times, Science Section
11894:from the original on 28 August 2018
10879:
10633:
9806:. Selected Papers from IJCNN 2011.
9795:
9225:
8811:
8688:
7554:
7510:
7449:
6787:; Osindero, S.; Teh, Y. W. (2006).
5198:. New York: John Wiley & Sons.
4843:"A Stochastic Approximation Method"
4739:. American Elsevier Publishing Co.
4259:Murphy, Kevin P. (24 August 2012).
3942:
3902:"Early History of Machine Learning"
3722:Shigeki, Sugiyama (12 April 2019).
2840:Explainable artificial intelligence
2678:
2499:, such that the conditions for the
1989:Triphone GMM-HMM with BMMI Training
914:is an efficient application of the
440:. The field takes inspiration from
24:
14012:
13739:
13527:Marcus, Gary (November 25, 2012).
13174:from the original on 28 March 2014
10651:from the original on 30 April 2017
10323:from the original on 13 March 2017
7250:Seide, F.; Li, G.; Yu, D. (2011).
7189:
7168:from the original on 16 March 2018
6964:Hinton, Geoffrey E. (2009-05-31).
6464:. McGill University Ph.D. thesis.
6118:
5175:from the original on 14 April 2016
3172:Facebook's face recognition system
2956:Another group showed that certain
2800:system, which learned the game of
2247:probabilistic context free grammar
1973:Monophone Randomly Initialized DNN
1491:
665:
60:
25:
15057:
13596:from the original on 19 June 2015
11719:Global Banking and Finance Review
11281:"Neural Joint Entropy Estimation"
10511:Arrowsmith, J; Miller, P (2013).
10392:from the original on 10 July 2017
10354:from the original on 7 April 2017
10267:Journal of Translational Medicine
9268:from the original on 17 June 2020
9207:from the original on 10 June 2020
9063:from the original on 29 July 2020
8643:Mikolov, T.; et al. (2010).
6065:(1995). "The Helmholtz machine".
5005:. Open court publishing Company.
3474:from the original on 17 June 2018
3363:
2816:The University of Texas at Austin
2781:'s AI lab performs tasks such as
2581:In November 2023, researchers at
2483:tasks, the NJEE maps a vector of
2402:
2284:example-based machine translation
1053:. Hochreiter proposed recurrent
968:recognizing handwritten ZIP codes
954:(TDNN) was introduced in 1987 by
559:, although they can also include
15010:
15009:
14989:
13971:
13896:
13858:
13789:
13759:
13733:
13724:
13696:Found. Trends Comput. Graph. Vis
13687:
13666:
13645:
13577:
13565:from the original on 3 July 2015
13550:
13520:
13465:Marcus, Gary (14 January 2018).
13458:
13428:
13387:
13375:from the original on 7 July 2019
13356:
13326:
13318:
13243:
13186:
13096:
13031:
12980:
12929:
12872:
12797:
12746:
12679:
12636:
12569:
12369:
12344:
12276:
12211:
12154:
12028:
11971:
11919:Journal of Computational Physics
11906:
11828:
11761:
11736:
11706:
11671:
11637:
11607:
11566:
11507:
11450:
11394:
11332:
11245:
11221:
11164:
11105:
11042:
11011:
10958:
10924:
10903:
10873:
10830:
10809:
10779:
10723:
10495:
10433:Metz, Cade (27 September 2016).
10305:
10254:
10115:
10084:
10011:
9999:from the original on 6 July 2014
9870:(6 January 2015), David Talbot,
9861:
9732:from the original on 8 June 2017
9713:
9548:
9366:
9341:
9298:
9280:
9250:
9238:from the original on 25 May 2020
9219:
9189:
9165:Research, AI (23 October 2015).
9158:
9075:
9016:
8956:
8859:
8437:
8407:
8377:
8291:
8252:
8228:
8095:
8018:
7992:
7889:
7770:
7680:"Going deeper with convolutions"
7412:
7289:
7243:
7150:
7015:(2016). Slides on Deep Learning
6957:
6910:
6871:
6110:
5583:Elman, Jeffrey L. (March 1990).
5166:System modeling and optimization
4136:Orhan, A. E.; Ma, W. J. (2017).
3855:. Science & Business Media.
3013:
2718:spinoff company Deep Longevity.
2655:Physics-informed neural networks
2653:In addition, the integration of
2365:Customer relationship management
2359:Customer relationship management
2053:Sequence discriminative training
1530:
1506:
896:Deep learning architectures for
750:cumulative distribution function
444:and is centered around stacking
14085:from the original on 2016-04-16
13885:from the original on 2017-10-10
13634:from the original on 2015-05-13
13539:from the original on 2009-11-27
12938:Current Opinion in Neurobiology
11725:from the original on 2018-11-16
11660:from the original on 2018-01-02
11439:from the original on 2021-05-09
11093:from the original on 9 May 2021
10947:from the original on 2017-05-16
10892:from the original on 2020-04-30
10686:from the original on 2015-09-08
10622:from the original on 2020-07-16
10036:from the original on 2014-11-27
9784:from the original on 2014-01-13
9702:from the original on 2020-09-24
9613:from the original on 2020-09-22
9005:from the original on 2021-05-09
8945:from the original on 2017-08-12
8779:LeCun, Y.; et al. (1998).
8760:from the original on 9 May 2021
8677:from the original on 2017-05-16
8577:from the original on 2021-05-09
8518:from the original on 2020-01-26
8443:
8426:from the original on 2021-01-07
8396:from the original on 2017-06-29
8367:from the original on 2020-11-02
8217:from the original on 2021-05-09
8199:Zen, Heiga; Sak, Hasim (2015).
8122:10.1109/ICCCI50826.2021.9402569
8084:from the original on 2016-03-09
7920:Generative Adversarial Networks
7577:from the original on 2017-08-09
7543:from the original on 2014-09-29
7378:from the original on 2020-05-18
7278:from the original on 2017-10-12
7232:from the original on 2017-10-12
7202:from the original on 2017-09-26
7106:from the original on 2017-09-26
7046:IEEE Signal Processing Magazine
6835:from the original on 2015-12-23
6710:
6693:
6630:from the original on 2021-05-09
6590:from the original on 9 May 2021
6572:
6563:
6468:from the original on 2021-05-09
6411:IEEE Signal Processing Magazine
6391:from the original on 2021-04-27
6330:
6319:from the original on 2021-05-09
6252:
6205:
6187:Sejnowski, Terrence J. (2018).
6180:
6005:
5952:
5770:
5576:
5553:
5447:
5400:
5212:
5129:"Who Invented Backpropagation?"
4889:
4820:from the original on 2017-08-29
4760:Ivakhnenko, A.G. (March 1970).
4753:
4643:
4387:from the original on 2017-01-11
4361:Bishop, Christopher M. (2006).
4234:from the original on 2016-03-14
4129:
3893:
3884:
3765:from the original on 2019-10-20
3715:
3445:from the original on 2017-01-10
3024:needs additional citations for
2932:
2911:artificial general intelligence
2501:universal approximation theorem
2447:Deep Neural Network Estimations
2260:Recent developments generalize
2069:by DNNs and related deep models
1896:
1893:in data-heavy AI applications.
1157:, SRI researched in speech and
1086:generative adversarial networks
672:universal approximation theorem
589:model, the raw input may be an
482:generative adversarial networks
81:Artificial general intelligence
30:For the TV series episode, see
14922:Recurrent neural network (RNN)
14912:Differentiable neural computer
13496:Knight, Will (14 March 2017).
13363:Metz, Cade (6 November 2017).
13131:10.1523/jneurosci.5023-14.2015
11713:Czech, Tomasz (28 June 2018).
9519:. Linguistic Data Consortium.
9033:. SC '17, ACM. pp. 1–12.
8927:Dahl, G.; et al. (2013).
8210:. ICASSP. pp. 4470–4474.
5286:Waibel, Alex (December 1987).
5168:. Springer. pp. 762–770.
3486:
3456:
3255:
2621:partial differential equations
2615:Partial differential equations
1402:Generative adversarial network
945:
784:
13:
1:
14967:Variational autoencoder (VAE)
14927:Long short-term memory (LSTM)
14194:Computational learning theory
13162:Metz, C. (12 December 2013).
12898:10.1016/s0896-6273(00)81098-3
12512:Behavioral and Brain Sciences
11743:Nuñez, Michael (2023-11-29).
10517:Nature Reviews Drug Discovery
9641:10.21437/Interspeech.2014-433
9084:The Journal of Supercomputing
8660:10.21437/Interspeech.2010-343
7264:10.21437/Interspeech.2011-169
6557:10.1016/s0167-6393(99)00077-1
6527:10.1016/S0167-6393(99)00080-1
5977:10.1016/S0364-0213(85)80012-4
3919:10.1016/j.ifacol.2020.12.1888
3249:
3229:Scale space and deep learning
2820:U.S. Army Research Laboratory
2376:actions, defined in terms of
2300:Drug discovery and toxicology
2274:(GT) uses a large end-to-end
2124:voice search, and a range of
1695:
1686:Convolutional neural networks
1634:playing board and video games
1043:refers to such pre-training.
898:convolutional neural networks
848:Group method of data handling
553:convolutional neural networks
478:convolutional neural networks
14947:Convolutional neural network
14039:Prince, Simon J. D. (2023).
12831:10.1371/journal.pcbi.1002211
12489:10.1016/0893-6080(96)00033-0
11692:10.1016/j.knosys.2019.105048
11644:Schmidt, Uwe; Roth, Stefan.
10579:10.1016/j.drudis.2014.12.014
9826:10.1016/j.neunet.2012.02.023
7605:10.1007/978-3-642-40763-5_51
7352:10.1016/j.patcog.2004.01.013
6884:Trends in Cognitive Sciences
6866:Trends in Cognitive Sciences
6723:Trends in Cognitive Sciences
6500:10.1016/0893-6080(94)90027-2
6230:10.1016/0022-2836(88)90564-5
6218:Journal of Molecular Biology
6189:The deep learning revolution
5930:10.1016/j.neunet.2020.04.008
5697:Schmidhuber, Jürgen (1993).
5655:Schmidhuber, Jürgen (1992).
4778:10.1016/0005-1098(70)90092-0
4589:10.1007/978-1-4615-7566-5_15
4086:Hassoun, Mohamad H. (1995).
4032:10.1016/0893-6080(91)90009-t
3908:. 21st IFAC World Congress.
3693:10.1016/j.neunet.2014.09.003
2437:predicting protein structure
1902:Automatic speech recognition
1452:automatic speech recognition
1245:, but convergence was slow.
1121:protein structure prediction
1109:restricted Boltzmann machine
978:-5 (1998), a 7-level CNN by
962:et al. created a CNN called
719:Lebesgue integrable function
7:
14942:Multilayer perceptron (MLP)
14065:; Courville, Aaron (2016).
14041:Understanding deep learning
13404:10.1109/devlrn.2008.4640845
11835:Peplow, Mark (2023-11-29).
11583:10.1109/CSCITA.2017.8066548
11484:10.1109/ACCESS.2020.3006362
11372:10.1016/j.media.2017.07.005
9568:10.13140/RG.2.2.15418.90567
9348:Woodie, Alex (2021-11-01).
8999:Tech. Rep. UTML TR 2010-003
8896:10.1109/icassp.2013.6639349
8835:10.1109/icassp.2013.6639347
7678:Szegedy, Christian (2015).
7190:Li, Deng (September 2014).
6935:10.1162/neco.2006.18.7.1527
6810:10.1162/neco.2006.18.7.1527
5585:"Finding Structure in Time"
5560:Jordan, Michael I. (1986).
4794:Ivakhnenko, Alexey (1971).
4705:Principles of Neurodynamics
3268:KI - Künstliche Intelligenz
3177:
2814:As of 2008, researchers at
2606:
2370:Deep reinforcement learning
2224:Natural language processing
2218:Natural language processing
1803:
1589:, (analogous to biological
1031:network into a lower level
876:stochastic gradient descent
864:stochastic gradient descent
683:feedforward neural networks
542:
502:natural language processing
116:Natural language processing
10:
15062:
15018:Artificial neural networks
14932:Gated recurrent unit (GRU)
14158:Differentiable programming
14102:: CS1 maint: postscript (
13865:Gibney, Elizabeth (2017).
13584:Alex Hern (18 June 2015).
13193:Gibney, Elizabeth (2016).
12950:10.1016/j.conb.2004.07.007
12810:PLOS Computational Biology
12448:Elman, Jeffrey L. (1998).
12246:10.1109/TUFFC.2020.3010186
11849:10.1038/d41586-023-03745-5
11795:10.1038/s41586-023-06735-9
11310:10.1109/TNNLS.2022.3204919
10280:10.1186/s12967-023-04011-y
10171:10.1109/taslp.2014.2383614
9599:10.1109/taslp.2014.2339736
9484:10.1038/s41586-020-03070-1
8709:10.1162/neco.1997.9.8.1735
8261:Neural Networks for Babies
6896:10.1016/j.tics.2007.09.004
6735:10.1016/j.tics.2007.09.004
5632:"Neural Sequence Chunkers"
5601:10.1207/s15516709cog1402_1
4340:10.1016/j.acha.2015.12.005
4163:10.1038/s41467-017-00181-8
3199:Differentiable programming
2837:
2682:
2406:
2391:
2362:
2304:For more information, see
2303:
2235:. Word embedding, such as
2221:
2135:
2085:and its rich LSTM variants
1997:Monophone DBN-DNN on fbank
1905:
1571:biological neural networks
1556:Artificial neural networks
1495:
1051:vanishing gradient problem
795:feedforward neural network
779:
626:feedforward neural network
524:, material inspection and
169:Hybrid intelligent systems
91:Recursive self-improvement
32:Deep Learning (South Park)
29:
27:Branch of machine learning
14985:
14899:
14843:
14772:
14705:
14577:
14477:
14470:
14424:
14388:
14351:Artificial neural network
14331:
14207:
14174:Automatic differentiation
14147:
13879:10.1038/nature.2017.22784
12767:10.1038/s41562-017-0186-2
12657:10.1162/neco.1996.8.5.895
12534:10.1017/s0140525x97001581
12419:10.1162/08997660260293319
12005:10.1016/j.cma.2019.112789
11940:10.1016/j.jcp.2018.10.045
11252:Shead, Sam (2020-11-30).
11055:. ACM. pp. 533–540.
10851:10.1038/s41587-019-0224-x
9684:10.1186/s13636-015-0068-3
9424:10.1038/s41586-020-2861-0
9114:10.1007/s11227-017-1994-x
7220:Yu, D.; Deng, L. (2010).
7130:Yu, D.; Deng, L. (2014).
6991:10.4249/scholarpedia.5947
6273:10.1142/s0218001493000455
6088:21.11116/0000-0002-D6D3-E
6079:10.1162/neco.1995.7.5.889
5879:10.1109/TAMD.2010.2056368
5756:. John Wiley & Sons.
5676:10.1162/neco.1992.4.2.234
5079:BIT Numerical Mathematics
4919:Trans. IECE (In Japanese)
4811:10.1109/TSMC.1971.4308320
4622:10.1109/TSMC.1972.4309133
4448:10.1103/RevModPhys.39.883
4428:Reviews of Modern Physics
4204:Deng, L.; Yu, D. (2014).
4092:. MIT Press. p. 48.
4056:Haykin, Simon S. (1999).
3861:10.1007/978-1-4757-3115-6
3813:10.4249/scholarpedia.5947
3393:10.1109/cvpr.2012.6248110
3280:10.1007/s13218-012-0198-z
3244:Topological deep learning
2962:facial recognition system
2899:
2833:
2732:cognitive neuroscientists
2689:An epigenetic clock is a
2565:Financial fraud detection
2157:remotely operated vehicle
1957:Bayesian Triphone GMM-HMM
1766:{\displaystyle \ell _{1}}
1735:{\displaystyle \ell _{2}}
1676:Recurrent neural networks
1498:Artificial neural network
1408:et al., 2014) (based on
1222:In 2006, publications by
986:Recurrent neural networks
952:time delay neural network
920:Gottfried Wilhelm Leibniz
866:was published in 1967 by
827:recurrent neural networks
803:recurrent neural networks
744:, related to fitting and
630:recurrent neural networks
474:recurrent neural networks
14179:Neuromorphic engineering
14142:Differentiable computing
14092:, introductory textbook.
13948:10.1177/1461444819885334
13561:. Google Research Blog.
12705:10.3389/fncom.2016.00073
12189:10.1109/TCI.2021.3075349
11118:JMIR mHealth and uHealth
7066:10.1109/msp.2012.2205597
6868:, 11, pp. 428–434, 2007.
6012:Smolensky, Paul (1986).
5194:Werbos, Paul J. (1994).
4520:Hopfield, J. J. (1982).
4303:10.1109/TSSC.1969.300225
2960:spectacles could fool a
2477:probability distribution
2430:electronic health record
2255:named-entity recognition
1949:Randomly Initialized RNN
1862:field-effect transistors
1827:deep learning processors
1662:multivariate polynomials
1265:Deep learning revolution
1187:
1078:probability distribution
1017:internal representations
1009:self-supervised learning
889:(rectified linear unit)
872:internal representations
858:The first deep learning
466:fully connected networks
293:Artificial consciousness
14952:Residual neural network
14368:Artificial Intelligence
13935:New Media & Society
13109:Journal of Neuroscience
12603:10.1073/pnas.88.10.4433
12321:10.1126/science.adi2336
12129:10.1073/pnas.1718942115
12061:10.1126/science.aaw4741
11680:Knowledge-Based Systems
11533:10.3390/cancers14071819
11061:10.1145/2649387.2649442
10645:www.datascienceassn.org
10471:"MT on and for the Web"
10348:The Keyword Google Blog
9987:"Deep Learning for NLP"
9334:10.1145/3140659.3080246
9039:10.1145/3126908.3126912
8785:Proceedings of the IEEE
7433:10.1145/1553374.1553486
6431:10.1109/msp.2009.932166
6158:10.1126/science.7761831
5979:(inactive 2024-08-07).
5510:Proceedings of the IEEE
5350:, 1, pp. 541–551, 1989.
4921:. J62-A (10): 658–665.
4860:10.1214/aoms/1177729586
4650:Rosenblatt, F. (1958).
4409:"bibliotheca Augustana"
3525:Bengio, Yoshua (2009).
3115:that not only low-paid
2927:artificial intelligence
2704:frontotemporal dementia
2625:Navier-Stokes equations
2382:customer lifetime value
2291:Forensic Identification
1839:tensor processing units
1785:, and initial weights.
1636:and medical diagnosis.
1381:residual neural network
1370:generating descriptions
1217:handwriting recognition
1198:support vector machines
1023:into a single RNN, by
676:probabilistic inference
442:biological neuroscience
438:representation learning
164:Evolutionary algorithms
54:Artificial intelligence
13903:Tubaro, Paola (2020).
13620:Goertzel, Ben (2015).
13440:governmentciomedia.com
13056:10.1098/rstb.2017.0043
13044:Phil. Trans. R. Soc. B
12755:Nature Human Behaviour
11342:Medical Image Analysis
9562:. CUED/F-INFENG/TR82.
8993:Hinton, G. E. (2010).
8870:. pp. 8624–8628.
8829:. pp. 8614–8618.
6966:"Deep belief networks"
6305:. Icassp'92: 617–620.
6025:. MIT Press. pp.
5787:Long Short Term Memory
4805:. SMC-1 (4): 364–378.
4718:Joseph, R. D. (1960).
4616:. SMC-2 (3): 380–388.
4547:10.1073/pnas.79.8.2554
3788:"Deep belief networks"
3377:. pp. 3642–3649.
3121:Amazon Mechanical Turk
3005:Data collection ethics
2886:
2511:Medical image analysis
2388:Recommendation systems
2276:long short-term memory
2196:
2163:
1813:commercial cloud AI .
1767:
1736:
1704:and computation time.
1580:rule-based programming
1546:result for sea urchin.
1484:were awarded the 2018
1301:Luca Maria Gambardella
1274:
1215:contest, in connected
1103:, etc., including the
1059:long short-term memory
841:multilayer perceptrons
622:credit assignment path
561:propositional formulas
518:medical image analysis
490:neural radiance fields
65:
42:
14907:Neural Turing machine
14495:Human image synthesis
13773:. 11 September 2017.
13502:MIT Technology Review
13338:MIT Technology Review
11415:10.1109/ICCVW.2017.18
11233:MIT Technology Review
9873:MIT Technology Review
9662:Tóth, Laszló (2015).
9226:Ray, Tiernan (2019).
8450:MIT Technology Review
5639:TR FKI-148, TU Munich
4142:Nature Communications
3786:Hinton, G.E. (2009).
3625:10.1109/tpami.2013.50
2919:commonsense reasoning
2862:
2826:Criticism and comment
2790:DeepMind Technologies
2352:graph neural networks
2208:Neural Style Transfer
2190:
2183:Visual art processing
2154:
1944:error rate (PER) (%)
1843:Google Cloud Platform
1768:
1737:
1392:neural style transfer
1295:In 2011, a CNN named
1272:
860:multilayer perceptron
825:in 1982. Other early
799:multilayer perceptron
704:rectified linear unit
567:such as the nodes in
64:
40:
14998:Computer programming
14977:Graph neural network
14552:Text-to-video models
14530:Text-to-image models
14378:Large language model
14363:Scientific computing
14169:Statistical manifold
14164:Information geometry
13398:. pp. 292–297.
11621:. 13 November 2018.
11577:. pp. 174–177.
11189:10.1093/jamia/ocw112
11131:10.2196/mhealth.6562
10839:Nature Biotechnology
10566:Drug Discovery Today
10386:Google Research Blog
8823:Ramabhadran, Bhuvana
7873:10.1109/CVPR.2016.90
7480:10.1162/neco_a_00052
7258:. pp. 437–440.
6623:. pp. 175–184.
6545:Speech Communication
6515:Speech Communication
5433:10.1364/AO.30.004211
5386:10.1364/AO.29.004790
5125:Schmidhuber, Juergen
4841:; Monro, S. (1951).
4707:. Spartan, New York.
4656:Psychological Review
4583:. pp. 172–186.
3214:Liquid state machine
3155:(e.g. by leveraging
3141:search results pages
3033:improve this article
2866:causal relationships
2764:deep belief networks
2661:Image reconstruction
2648:deep neural networks
2505:consistent estimator
2497:activation functions
2481:image classification
2337:rational drug design
2322:biomolecular targets
2021:Ensemble DNN/CNN/RNN
1882:in conjunction with
1872:hardware accelerator
1750:
1742:-regularization) or
1719:
1644:Deep neural networks
1597:). Each connection (
1464:image classification
1347:ImageNet competition
1345:won the large-scale
1235:deep belief networks
1228:Ruslan Salakhutdinov
1117:wake-sleep algorithm
998:cognitive psychology
711:deep neural networks
687:continuous functions
645:deep belief networks
569:deep belief networks
470:deep belief networks
106:General game playing
18:Deep neural networks
14344:In-context learning
14184:Pattern recognition
13842:. 10 October 2017.
13289:10.1038/nature16961
13281:2016Natur.529..484S
13211:2016Natur.529..445G
13115:(27): 10005–10014.
12989:Nature Neuroscience
12822:2011PLSCB...7E2211B
12594:1991PNAS...88.4433M
12312:2023Sci...382.1416L
12296:(6677): 1416–1421.
12120:2018PNAS..115.8505H
12053:2020Sci...367.1026R
12047:(6481): 1026–1030.
11996:2020CMAME.360k2789M
11931:2019JCoPh.378..686R
11786:2023Natur.624...80M
11475:2020IEEEA...8l9889D
11364:2017arXiv170205747L
10711:on 28 February 2015
9947:10.3390/arts6040018
9909:10.3390/arts6020005
9416:2020Natur.587...72M
9292:consumer.huawei.com
9106:2017arXiv170207908V
8568:2014arXiv1409.3215S
8321:10.1038/nature16961
8313:2016Natur.529..484S
8040:. PMLR: 2256–2265.
7959:. December 14, 2018
7344:2004PatRe..37.1311O
7332:Pattern Recognition
7164:. 3 December 2015.
7058:2012ISPM...29...82H
6982:2009SchpJ...4.5947H
6653:Schmidhuber, Jürgen
6460:Bengio, Y. (1991).
6423:2009ISPM...26...75B
6150:1995Sci...268.1158H
6144:(5214): 1158–1161.
6126:Hinton, Geoffrey E.
6055:Hinton, Geoffrey E.
5904:Schmidhuber, Jürgen
5863:Schmidhuber, Jürgen
5845:Schmidhuber, Jürgen
5818:10.1049/cp:19991218
5628:Schmidhuber, Jürgen
5472:1994MedPh..21..517Z
5454:Zhang, Wei (1994).
5425:1991ApOpt..30.4211Z
5407:Zhang, Wei (1991).
5378:1990ApOpt..29.4790Z
5360:Zhang, Wei (1990).
5324:Zhang, Wei (1988).
5237:1986Natur.323..533R
4538:1982PNAS...79.2554H
4497:Schmidhuber, Jürgen
4440:1967RvMP...39..883B
4154:2017NatCo...8..138O
4002:on 10 October 2015.
3977:1989MCSS....2..303C
3804:2009SchpJ...4.5947H
3341:10.1038/nature14539
3333:2015Natur.521..436L
3224:Reservoir computing
2882:deductive reasoning
2774:Commercial activity
2737:nerve growth factor
2631:methods relies on.
2467:that maps an input
1823:electronic circuits
1626:machine translation
1441:Google Voice Search
1213:pattern recognition
1159:speaker recognition
1136:Hidden Markov model
891:activation function
610:feature engineering
535:, particularly the
506:machine translation
258:Machine translation
174:Systems integration
111:Knowledge reasoning
48:Part of a series on
14937:Echo state network
14825:Jürgen Schmidhuber
14520:Facial recognition
14515:Speech recognition
14425:Software libraries
13718:10.1561/0600000018
13369:The New York Times
13344:on 1 February 2016
13050:(1740): 20170043.
12645:Neural Computation
12566:, pp. B5–B6, 1995.
12407:Neural Computation
11409:. pp. 82–89.
11024:Microsoft Research
10791:KQED Future of You
10760:The Globe and Mail
10317:Microsoft Research
10201:Microsoft Research
10128:Microsoft Research
10097:Microsoft Research
9525:10.35111/17gk-bn40
9151:2018-11-18 at the
9143:Ting Qin, et al. "
8697:Neural Computation
7458:Neural Computation
7302:Microsoft Research
7162:Microsoft Research
7022:2016-04-23 at the
6923:Neural Computation
6860:2018-05-22 at the
6797:Neural Computation
6686:2018-11-18 at the
6366:10338.dmlcz/135496
6067:Neural Computation
5784:(21 August 1995),
5782:Jürgen Schmidhuber
5730:2015-03-06 at the
5664:Neural Computation
5348:Neural Computation
5272:2022-10-13 at the
5091:10.1007/bf01931367
4970:10.1007/bf00344251
4927:10.1007/bf00344251
4413:www.hs-augsburg.de
4225:10.1561/2000000039
4122:2019-02-13 at the
3985:10.1007/bf02551274
3836:2016-04-19 at the
3556:10.1561/2200000006
3204:Echo state network
3194:Compressed sensing
3153:information mining
2940:adversarial attack
2915:Learning a grammar
2878:Bayesian inference
2870:logical inferences
2670:Weather prediction
2596:crystal structures
2526:mobile advertising
2520:Mobile advertising
2459:on an independent
2453:stochastic process
2394:Recommender system
2345:multiple sclerosis
2266:sentence embedding
2251:sentiment analysis
2197:
2164:
2161:mussel aquaculture
2029:Bidirectional LSTM
1908:Speech recognition
1763:
1732:
1622:speech recognition
1587:artificial neurons
1410:Jürgen Schmidhuber
1305:Jürgen Schmidhuber
1275:
1143:speech recognition
1128:speech recognition
1066:Jürgen Schmidhuber
1005:Jürgen Schmidhuber
940:David E. Rumelhart
906:Kunihiko Fukushima
883:Kunihiko Fukushima
700:Kunihiko Fukushima
593:(represented as a
573:Boltzmann machines
533:biological systems
498:speech recognition
446:artificial neurons
66:
43:
15033:
15032:
14795:Stephen Grossberg
14768:
14767:
14043:. The MIT Press.
14031:978-3-031-45467-7
13941:(10): 1868–1884.
13413:978-1-4244-2661-4
13275:(7587): 484–489.
13205:(7587): 445–446.
12588:(10): 4433–4437.
12461:978-0-262-55030-7
12413:(10): 2497–2529.
12382:Aging and Disease
12230:(12): 2584–2594.
12104:(34): 8505–8510.
11592:978-1-5090-4381-1
11469:: 129889–129898.
10880:Gregory, Barber.
9633:Proc. Interspeech
9593:(10): 1533–1545.
9203:. December 2019.
8905:978-1-4799-0356-6
8844:978-1-4799-0356-6
8791:(11): 2278–2324.
8496:10.1109/72.963769
8307:(7587): 484–489.
8167:on 24 April 2018.
8131:978-1-7281-5875-4
7882:978-1-4673-8851-1
7614:978-3-642-38708-1
7464:(12): 3207–3220.
7442:978-1-60558-516-1
7143:978-1-4471-5779-3
6198:978-0-262-03803-4
6063:Zemel, Richard S.
5965:Cognitive Science
5763:978-0-7803-5369-5
5722:S. Hochreiter., "
5589:Cognitive Science
5516:(11): 2278–2324.
5231:(6088): 533–536.
5075:Linnainmaa, Seppo
5057:Linnainmaa, Seppo
4879:IEEE Transactions
4746:978-0-444-00020-0
4701:Rosenblatt, Frank
4598:978-1-4615-7568-9
4466:IEEE Transactions
4377:978-0-387-31073-2
4272:978-0-262-01802-9
4099:978-0-262-08239-6
4069:978-0-13-273350-2
4062:. Prentice Hall.
3906:IFAC-PapersOnLine
3870:978-0-7923-7824-2
3735:978-1-5225-8218-2
3402:978-1-4673-1228-8
3327:(7553): 436–444.
3239:Stochastic parrot
3161:activity trackers
3109:
3108:
3101:
3083:
2985:genetic algorithm
2760:generative models
2741:self-organization
2728:brain development
2716:Insilico Medicine
2600:Materials Project
2591:materials science
2577:Materials science
2555:film colorization
2533:Image restoration
2152:
2132:Image recognition
2067:transfer learning
2044:
2043:
2005:Convolutional DNN
1981:Monophone DBN-DNN
1810:computer hardware
1690:acoustic modeling
1680:language modeling
1329:In October 2012,
1147:SRI International
1113:Helmholtz machine
1105:Boltzmann machine
1013:predictive coding
972:optical computing
852:Alexey Ivakhnenko
587:image recognition
565:generative models
432:methods based on
423:
422:
159:Bayesian networks
86:Intelligent agent
16:(Redirected from
15053:
15023:Machine learning
15013:
15012:
14993:
14748:Action selection
14738:Self-driving car
14545:Stable Diffusion
14510:Speech synthesis
14475:
14474:
14339:Machine learning
14215:Gradient descent
14136:
14129:
14122:
14113:
14112:
14107:
14101:
14093:
14091:
14090:
14078:978-0-26203561-3
14054:
14035:
14007:
14006:
14004:
14002:
13975:
13969:
13968:
13950:
13926:
13913:
13912:
13900:
13894:
13893:
13891:
13890:
13862:
13856:
13855:
13853:
13851:
13832:
13817:
13816:
13814:
13812:
13803:. 18 June 2018.
13793:
13787:
13786:
13784:
13782:
13763:
13757:
13756:
13754:
13753:
13744:. Archived from
13737:
13731:
13728:
13722:
13721:
13711:
13691:
13685:
13684:
13682:
13670:
13664:
13663:
13661:
13649:
13643:
13642:
13640:
13639:
13633:
13626:
13617:
13606:
13605:
13603:
13601:
13581:
13575:
13574:
13572:
13570:
13554:
13548:
13547:
13545:
13544:
13524:
13518:
13517:
13515:
13513:
13493:
13487:
13486:
13484:
13482:
13462:
13456:
13455:
13453:
13451:
13432:
13426:
13425:
13391:
13385:
13384:
13382:
13380:
13360:
13354:
13353:
13351:
13349:
13340:. Archived from
13330:
13324:
13323:
13322:
13316:
13247:
13241:
13240:
13222:
13190:
13184:
13183:
13181:
13179:
13159:
13153:
13152:
13142:
13124:
13100:
13094:
13093:
13075:
13035:
13029:
13028:
12984:
12978:
12977:
12933:
12927:
12926:
12900:
12876:
12870:
12869:
12851:
12833:
12816:(11): e1002211.
12801:
12795:
12794:
12750:
12744:
12743:
12725:
12707:
12683:
12677:
12676:
12640:
12634:
12633:
12623:
12605:
12573:
12567:
12560:
12554:
12553:
12527:
12507:
12501:
12500:
12483:(7): 1119–1129.
12472:
12466:
12465:
12445:
12439:
12438:
12402:
12396:
12395:
12393:
12373:
12367:
12366:
12364:
12363:
12348:
12342:
12341:
12323:
12305:
12280:
12274:
12273:
12239:
12215:
12209:
12208:
12182:
12158:
12152:
12151:
12141:
12131:
12113:
12089:
12083:
12082:
12072:
12032:
12026:
12025:
12007:
11975:
11969:
11968:
11942:
11910:
11904:
11903:
11901:
11899:
11880:
11869:
11868:
11832:
11826:
11825:
11815:
11797:
11765:
11759:
11758:
11756:
11755:
11740:
11734:
11733:
11731:
11730:
11710:
11704:
11703:
11675:
11669:
11668:
11666:
11665:
11659:
11652:
11641:
11635:
11634:
11632:
11630:
11611:
11605:
11604:
11570:
11564:
11563:
11553:
11535:
11511:
11505:
11504:
11486:
11454:
11448:
11447:
11445:
11444:
11398:
11392:
11391:
11357:
11336:
11330:
11329:
11303:
11294:(4): 5488–5500.
11285:
11276:
11267:
11266:
11264:
11263:
11249:
11243:
11242:
11240:
11239:
11225:
11219:
11218:
11208:
11168:
11162:
11161:
11143:
11133:
11109:
11103:
11102:
11100:
11098:
11046:
11040:
11039:
11037:
11035:
11015:
11009:
11008:
10998:
10988:
10962:
10956:
10955:
10953:
10952:
10946:
10939:
10928:
10922:
10921:
10919:
10907:
10901:
10900:
10898:
10897:
10877:
10871:
10870:
10845:(9): 1038–1040.
10834:
10828:
10827:
10825:
10813:
10807:
10806:
10804:
10802:
10783:
10777:
10776:
10774:
10772:
10751:
10742:
10741:
10739:
10727:
10721:
10720:
10718:
10716:
10707:. Archived from
10701:
10695:
10694:
10692:
10691:
10676:
10670:
10667:
10661:
10660:
10658:
10656:
10637:
10631:
10630:
10628:
10627:
10608:
10602:
10601:
10591:
10581:
10557:
10551:
10550:
10532:
10508:
10502:
10499:
10493:
10492:
10490:
10488:
10483:on 29 March 2017
10482:
10476:. Archived from
10475:
10466:
10457:
10456:
10454:
10452:
10430:
10424:
10423:
10421:
10408:
10402:
10401:
10399:
10397:
10377:
10364:
10363:
10361:
10359:
10339:
10333:
10332:
10330:
10328:
10309:
10303:
10302:
10292:
10282:
10258:
10252:
10251:
10240:10.1002/dac.3259
10223:
10217:
10216:
10214:
10212:
10192:
10183:
10182:
10150:
10144:
10143:
10141:
10139:
10119:
10113:
10112:
10110:
10108:
10088:
10082:
10081:
10079:
10077:
10071:
10060:
10051:
10045:
10044:
10042:
10041:
10035:
10024:
10015:
10009:
10008:
10006:
10004:
9998:
9991:
9982:
9973:
9972:
9970:
9958:
9952:
9951:
9949:
9925:
9914:
9913:
9911:
9887:
9876:
9865:
9859:
9858:
9856:
9844:
9838:
9837:
9819:
9799:
9793:
9792:
9790:
9789:
9770:
9764:
9763:
9761:
9748:
9742:
9741:
9739:
9737:
9717:
9711:
9710:
9708:
9707:
9701:
9686:
9668:
9659:
9653:
9652:
9628:
9622:
9621:
9619:
9618:
9578:
9572:
9571:
9552:
9546:
9545:
9543:
9541:
9510:
9504:
9503:
9477:
9457:
9446:
9445:
9435:
9395:
9389:
9388:
9386:
9385:
9370:
9364:
9363:
9361:
9360:
9345:
9339:
9338:
9336:
9326:
9302:
9296:
9295:
9284:
9278:
9277:
9275:
9273:
9258:"AI and Compute"
9254:
9248:
9247:
9245:
9243:
9223:
9217:
9216:
9214:
9212:
9193:
9187:
9186:
9184:
9182:
9162:
9156:
9141:
9135:
9132:
9126:
9125:
9099:
9079:
9073:
9072:
9070:
9068:
9020:
9014:
9013:
9011:
9010:
8990:
8984:
8983:
8981:
8979:
8960:
8954:
8953:
8951:
8950:
8944:
8933:
8924:
8918:
8917:
8889:
8879:
8863:
8857:
8856:
8819:Sainath, Tara N.
8815:
8809:
8808:
8797:10.1109/5.726791
8776:
8770:
8769:
8767:
8765:
8746:
8737:
8736:
8703:(8): 1735–1780.
8692:
8686:
8685:
8683:
8682:
8676:
8649:
8640:
8634:
8633:
8631:
8619:
8610:
8609:
8607:
8595:
8586:
8585:
8583:
8582:
8576:
8561:
8547:
8538:
8527:
8526:
8524:
8523:
8490:(6): 1333–1340.
8475:
8466:
8465:
8463:
8461:
8456:on 31 March 2019
8452:. Archived from
8441:
8435:
8434:
8432:
8431:
8411:
8405:
8404:
8402:
8401:
8381:
8375:
8374:
8373:
8372:
8355:
8349:
8348:
8295:
8289:
8288:
8282:
8274:
8256:
8250:
8249:
8247:
8246:
8232:
8226:
8225:
8223:
8222:
8216:
8205:
8196:
8190:
8189:
8187:
8175:
8169:
8168:
8166:
8160:. Archived from
8159:
8150:
8144:
8143:
8115:
8106:. pp. 1–4.
8099:
8093:
8092:
8090:
8089:
8073:
8064:
8058:
8052:
8051:
8049:
8031:
8022:
8016:
8015:
8013:
8011:
7996:
7990:
7989:
7987:
7975:
7969:
7968:
7966:
7964:
7957:SyncedReview.com
7949:
7943:
7942:
7940:
7938:
7932:
7925:
7914:
7908:
7907:
7905:
7893:
7887:
7886:
7866:
7844:
7838:
7837:
7835:
7819:
7813:
7812:
7810:
7798:
7792:
7791:
7790:
7774:
7768:
7766:
7764:
7752:
7746:
7744:
7742:
7729:
7723:
7721:
7719:
7707:
7701:
7700:
7698:
7684:
7675:
7669:
7668:
7666:
7654:
7648:
7647:
7645:
7633:
7627:
7626:
7592:
7586:
7585:
7583:
7582:
7576:
7569:
7558:
7552:
7551:
7549:
7548:
7542:
7523:
7514:
7508:
7507:
7473:
7453:
7447:
7446:
7416:
7410:
7409:
7407:
7392:
7386:
7385:
7384:
7383:
7365:
7356:
7355:
7338:(6): 1311–1314.
7327:
7318:
7317:
7315:
7313:
7293:
7287:
7286:
7284:
7283:
7256:Interspeech 2011
7247:
7241:
7240:
7238:
7237:
7217:
7211:
7210:
7208:
7207:
7187:
7178:
7177:
7175:
7173:
7154:
7148:
7147:
7127:
7116:
7115:
7113:
7111:
7105:
7098:
7089:
7078:
7077:
7037:
7026:
7010:
7004:
7003:
6993:
6961:
6955:
6954:
6929:(7): 1527–1554.
6914:
6908:
6907:
6875:
6869:
6852:G. E. Hinton., "
6850:
6844:
6843:
6841:
6840:
6834:
6804:(7): 1527–1554.
6793:
6781:
6775:
6774:
6772:
6770:
6714:
6708:
6697:
6691:
6677:
6671:
6670:
6668:
6645:
6639:
6638:
6636:
6635:
6629:
6618:
6609:
6600:
6599:
6597:
6595:
6576:
6570:
6567:
6561:
6560:
6540:
6531:
6530:
6510:
6504:
6503:
6483:
6477:
6476:
6474:
6473:
6457:
6451:
6450:
6406:
6400:
6399:
6397:
6396:
6390:
6358:10.1109/29.21701
6343:
6334:
6328:
6327:
6325:
6324:
6291:
6285:
6284:
6256:
6250:
6249:
6209:
6203:
6202:
6184:
6178:
6177:
6134:Frey, Brendan J.
6122:
6116:
6115:
6114:
6108:
6090:
6059:Neal, Radford M.
6047:
6041:
6040:
6018:
6009:
6003:
6002:
5996:
5988:
5956:
5950:
5949:
5923:
5900:
5891:
5890:
5859:
5853:
5852:
5841:
5832:
5831:
5805:
5799:
5798:
5774:
5768:
5767:
5745:
5739:
5720:
5709:
5707:
5705:
5694:
5688:
5687:
5661:
5652:
5643:
5642:
5636:
5624:
5613:
5612:
5580:
5574:
5573:
5557:
5551:
5550:
5548:
5546:
5532:10.1109/5.726791
5525:
5507:
5498:
5492:
5491:
5480:10.1118/1.597177
5451:
5445:
5444:
5404:
5398:
5397:
5357:
5351:
5340:
5334:
5333:
5321:
5315:
5306:Alexander Waibel
5303:
5297:
5296:
5294:
5283:
5277:
5263:
5257:
5256:
5245:10.1038/323533a0
5216:
5210:
5209:
5191:
5185:
5184:
5182:
5180:
5174:
5163:
5151:
5145:
5144:
5142:
5140:
5121:
5112:
5109:
5103:
5102:
5071:
5065:
5064:
5053:
5047:
5046:
5027:Kelley, Henry J.
5023:
5017:
5016:
4996:
4990:
4989:
4953:
4947:
4946:
4914:
4908:
4907:
4905:
4893:
4887:
4886:
4875:Amari, Shun'ichi
4871:
4865:
4864:
4862:
4835:
4829:
4828:
4826:
4825:
4819:
4800:
4791:
4782:
4781:
4757:
4751:
4750:
4730:
4724:
4723:
4715:
4709:
4708:
4697:
4688:
4687:
4668:10.1037/h0042519
4647:
4641:
4640:
4632:
4626:
4625:
4609:
4603:
4602:
4576:
4570:
4569:
4559:
4549:
4532:(8): 2554–2558.
4517:
4511:
4510:
4508:
4493:
4474:
4473:
4472:(21): 1197–1206.
4461:
4452:
4451:
4423:
4417:
4416:
4405:
4396:
4395:
4393:
4392:
4386:
4369:
4358:
4352:
4351:
4333:
4313:
4307:
4306:
4286:
4277:
4276:
4256:
4243:
4242:
4240:
4239:
4233:
4210:
4201:
4186:
4185:
4175:
4165:
4133:
4127:
4113:
4104:
4103:
4083:
4074:
4073:
4053:
4044:
4043:
4015:
4004:
4003:
4001:
3995:. Archived from
3960:
3954:Cybenko (1989).
3951:
3940:
3939:
3921:
3912:(2): 1385–1390.
3897:
3891:
3888:
3882:
3881:
3879:
3877:
3846:
3840:
3824:
3818:
3817:
3815:
3783:
3774:
3773:
3771:
3770:
3764:
3757:
3746:
3740:
3739:
3719:
3713:
3712:
3686:
3666:
3645:
3644:
3618:
3609:(8): 1798–1828.
3598:
3583:
3582:
3580:
3578:
3572:
3566:. Archived from
3549:
3531:
3522:
3509:
3508:
3506:
3505:
3490:
3484:
3483:
3481:
3479:
3460:
3454:
3453:
3451:
3450:
3444:
3433:
3424:
3415:
3414:
3386:
3370:
3361:
3360:
3318:
3309:
3300:
3299:
3259:
3159:devices such as
3113:media philosophy
3104:
3097:
3093:
3090:
3084:
3082:
3041:
3017:
3009:
2923:production rules
2806:Google Translate
2691:biochemical test
2685:Epigenetic clock
2679:Epigenetic clock
2559:Deep Image Prior
2547:super-resolution
2539:inverse problems
2457:random variables
2417:ANN was used in
2374:direct marketing
2278:(LSTM) network.
2272:Google Translate
2153:
2102:Skype Translator
2077:domain knowledge
1936:
1935:
1922:American English
1852:Atomically thin
1847:Cerebras Systems
1837:servers such as
1772:
1770:
1769:
1764:
1762:
1761:
1741:
1739:
1738:
1733:
1731:
1730:
1595:biological brain
1534:
1510:
1434:Stable Diffusion
1426:Diffusion models
1388:Google DeepDream
1359:Andrew Zisserman
1082:gradient descent
1074:generative model
932:Seppo Linnainmaa
837:Frank Rosenblatt
730:machine learning
580:machine learning
430:machine learning
415:
408:
401:
322:Existential risk
144:Machine learning
45:
44:
21:
15061:
15060:
15056:
15055:
15054:
15052:
15051:
15050:
15036:
15035:
15034:
15029:
14981:
14895:
14861:Google DeepMind
14839:
14805:Geoffrey Hinton
14764:
14701:
14627:Project Debater
14573:
14471:Implementations
14466:
14420:
14384:
14327:
14269:Backpropagation
14203:
14189:Tensor calculus
14143:
14140:
14110:
14095:
14094:
14088:
14086:
14079:
14059:Goodfellow, Ian
14051:
14032:
14015:
14013:Further reading
14010:
14000:
13998:
13977:
13976:
13972:
13927:
13916:
13909:Global Dialogue
13901:
13897:
13888:
13886:
13863:
13859:
13849:
13847:
13840:Singularity Hub
13834:
13833:
13820:
13810:
13808:
13795:
13794:
13790:
13780:
13778:
13765:
13764:
13760:
13751:
13749:
13740:Eisner, Jason.
13738:
13734:
13729:
13725:
13709:10.1.1.681.2190
13692:
13688:
13671:
13667:
13650:
13646:
13637:
13635:
13631:
13624:
13618:
13609:
13599:
13597:
13582:
13578:
13568:
13566:
13555:
13551:
13542:
13540:
13525:
13521:
13511:
13509:
13494:
13490:
13480:
13478:
13463:
13459:
13449:
13447:
13442:. 16 May 2018.
13434:
13433:
13429:
13414:
13392:
13388:
13378:
13376:
13361:
13357:
13347:
13345:
13332:
13331:
13327:
13317:
13263:Hassabis, Demis
13259:Sutskever, Ilya
13248:
13244:
13220:10.1038/529445a
13191:
13187:
13177:
13175:
13160:
13156:
13101:
13097:
13036:
13032:
13001:10.1038/nn.4244
12985:
12981:
12934:
12930:
12877:
12873:
12802:
12798:
12751:
12747:
12684:
12680:
12641:
12637:
12574:
12570:
12561:
12557:
12508:
12504:
12477:Neural Networks
12473:
12469:
12462:
12446:
12442:
12403:
12399:
12374:
12370:
12361:
12359:
12349:
12345:
12281:
12277:
12216:
12212:
12159:
12155:
12090:
12086:
12033:
12029:
11976:
11972:
11911:
11907:
11897:
11895:
11882:
11881:
11872:
11833:
11829:
11780:(7990): 80–85.
11766:
11762:
11753:
11751:
11741:
11737:
11728:
11726:
11711:
11707:
11676:
11672:
11663:
11661:
11657:
11650:
11642:
11638:
11628:
11626:
11613:
11612:
11608:
11593:
11571:
11567:
11512:
11508:
11455:
11451:
11442:
11440:
11425:
11399:
11395:
11337:
11333:
11283:
11277:
11270:
11261:
11259:
11250:
11246:
11237:
11235:
11227:
11226:
11222:
11169:
11165:
11110:
11106:
11096:
11094:
11079:
11047:
11043:
11033:
11031:
11016:
11012:
10963:
10959:
10950:
10948:
10944:
10937:
10929:
10925:
10908:
10904:
10895:
10893:
10878:
10874:
10835:
10831:
10814:
10810:
10800:
10798:
10793:. 27 May 2015.
10785:
10784:
10780:
10770:
10768:
10753:
10752:
10745:
10728:
10724:
10714:
10712:
10703:
10702:
10698:
10689:
10687:
10678:
10677:
10673:
10668:
10664:
10654:
10652:
10639:
10638:
10634:
10625:
10623:
10610:
10609:
10605:
10558:
10554:
10530:10.1038/nrd4090
10509:
10505:
10500:
10496:
10486:
10484:
10480:
10473:
10467:
10460:
10450:
10448:
10431:
10427:
10409:
10405:
10395:
10393:
10378:
10367:
10357:
10355:
10340:
10336:
10326:
10324:
10311:
10310:
10306:
10259:
10255:
10224:
10220:
10210:
10208:
10193:
10186:
10155:Hakkani-Tur, D.
10151:
10147:
10137:
10135:
10120:
10116:
10106:
10104:
10089:
10085:
10075:
10073:
10069:
10058:
10052:
10048:
10039:
10037:
10033:
10022:
10016:
10012:
10002:
10000:
9996:
9989:
9983:
9976:
9959:
9955:
9926:
9917:
9888:
9879:
9866:
9862:
9845:
9841:
9817:10.1.1.226.8219
9804:Neural Networks
9800:
9796:
9787:
9785:
9772:
9771:
9767:
9749:
9745:
9735:
9733:
9718:
9714:
9705:
9703:
9699:
9666:
9660:
9656:
9629:
9625:
9616:
9614:
9579:
9575:
9553:
9549:
9539:
9537:
9535:
9511:
9507:
9458:
9449:
9396:
9392:
9383:
9381:
9372:
9371:
9367:
9358:
9356:
9346:
9342:
9303:
9299:
9286:
9285:
9281:
9271:
9269:
9264:. 16 May 2018.
9256:
9255:
9251:
9241:
9239:
9224:
9220:
9210:
9208:
9201:InformationWeek
9195:
9194:
9190:
9180:
9178:
9163:
9159:
9153:Wayback Machine
9142:
9138:
9133:
9129:
9080:
9076:
9066:
9064:
9049:
9021:
9017:
9008:
9006:
8991:
8987:
8977:
8975:
8962:
8961:
8957:
8948:
8946:
8942:
8931:
8925:
8921:
8906:
8887:10.1.1.752.9151
8864:
8860:
8845:
8816:
8812:
8777:
8773:
8763:
8761:
8748:
8747:
8740:
8693:
8689:
8680:
8678:
8674:
8647:
8641:
8637:
8620:
8613:
8596:
8589:
8580:
8578:
8574:
8545:
8539:
8530:
8521:
8519:
8476:
8469:
8459:
8457:
8444:Hof, Robert D.
8442:
8438:
8429:
8427:
8412:
8408:
8399:
8397:
8382:
8378:
8370:
8368:
8357:
8356:
8352:
8296:
8292:
8276:
8275:
8271:
8263:. Sourcebooks.
8257:
8253:
8244:
8242:
8234:
8233:
8229:
8220:
8218:
8214:
8203:
8197:
8193:
8176:
8172:
8164:
8157:
8151:
8147:
8132:
8100:
8096:
8087:
8085:
8074:
8067:
8059:
8055:
8029:
8023:
8019:
8009:
8007:
8002:. witness.org.
7998:
7997:
7993:
7976:
7972:
7962:
7960:
7951:
7950:
7946:
7936:
7934:
7930:
7923:
7915:
7911:
7894:
7890:
7883:
7845:
7841:
7820:
7816:
7799:
7795:
7775:
7771:
7753:
7749:
7730:
7726:
7708:
7704:
7682:
7676:
7672:
7655:
7651:
7634:
7630:
7615:
7593:
7589:
7580:
7578:
7574:
7567:
7559:
7555:
7546:
7544:
7540:
7521:
7515:
7511:
7454:
7450:
7443:
7417:
7413:
7393:
7389:
7381:
7379:
7366:
7359:
7328:
7321:
7311:
7309:
7294:
7290:
7281:
7279:
7248:
7244:
7235:
7233:
7218:
7214:
7205:
7203:
7188:
7181:
7171:
7169:
7156:
7155:
7151:
7144:
7128:
7119:
7109:
7107:
7103:
7096:
7090:
7081:
7038:
7029:
7024:Wayback Machine
7011:
7007:
6962:
6958:
6915:
6911:
6890:(10): 428–434.
6876:
6872:
6862:Wayback Machine
6851:
6847:
6838:
6836:
6832:
6791:
6782:
6778:
6768:
6766:
6729:(10): 428–434.
6715:
6711:
6698:
6694:
6688:Wayback Machine
6678:
6674:
6646:
6642:
6633:
6631:
6627:
6616:
6610:
6603:
6593:
6591:
6578:
6577:
6573:
6568:
6564:
6541:
6534:
6511:
6507:
6488:Neural Networks
6484:
6480:
6471:
6469:
6458:
6454:
6407:
6403:
6394:
6392:
6388:
6341:
6335:
6331:
6322:
6320:
6313:
6292:
6288:
6257:
6253:
6210:
6206:
6199:
6185:
6181:
6123:
6119:
6109:
6048:
6044:
6037:
6016:
6010:
6006:
5990:
5989:
5957:
5953:
5908:Neural Networks
5901:
5894:
5860:
5856:
5842:
5835:
5828:
5806:
5802:
5778:Sepp Hochreiter
5775:
5771:
5764:
5746:
5742:
5732:Wayback Machine
5721:
5712:
5703:
5695:
5691:
5659:
5653:
5646:
5634:
5625:
5616:
5581:
5577:
5558:
5554:
5544:
5542:
5505:
5499:
5495:
5460:Medical Physics
5452:
5448:
5405:
5401:
5358:
5354:
5341:
5337:
5322:
5318:
5304:
5300:
5292:
5284:
5280:
5274:Wayback Machine
5264:
5260:
5217:
5213:
5206:
5192:
5188:
5178:
5176:
5172:
5161:
5152:
5148:
5138:
5136:
5135:on 30 July 2024
5127:(25 Oct 2014).
5122:
5115:
5110:
5106:
5072:
5068:
5054:
5050:
5037:(10): 947–954.
5024:
5020:
5013:
4997:
4993:
4954:
4950:
4915:
4911:
4894:
4890:
4872:
4868:
4836:
4832:
4823:
4821:
4817:
4798:
4792:
4785:
4758:
4754:
4747:
4731:
4727:
4716:
4712:
4698:
4691:
4648:
4644:
4633:
4629:
4610:
4606:
4599:
4577:
4573:
4518:
4514:
4494:
4477:
4462:
4455:
4424:
4420:
4407:
4406:
4399:
4390:
4388:
4384:
4378:
4367:
4359:
4355:
4314:
4310:
4287:
4280:
4273:
4257:
4246:
4237:
4235:
4231:
4208:
4202:
4189:
4134:
4130:
4124:Wayback Machine
4114:
4107:
4100:
4084:
4077:
4070:
4054:
4047:
4020:Neural Networks
4016:
4007:
3999:
3958:
3952:
3943:
3898:
3894:
3889:
3885:
3875:
3873:
3871:
3847:
3843:
3838:Wayback Machine
3825:
3821:
3784:
3777:
3768:
3766:
3762:
3755:
3747:
3743:
3736:
3720:
3716:
3671:Neural Networks
3667:
3648:
3599:
3586:
3576:
3574:
3573:on 4 March 2016
3570:
3547:10.1.1.701.9550
3529:
3523:
3512:
3503:
3501:
3492:
3491:
3487:
3477:
3475:
3470:. 25 May 2017.
3462:
3461:
3457:
3448:
3446:
3442:
3431:
3425:
3418:
3403:
3371:
3364:
3316:
3314:"Deep Learning"
3310:
3303:
3264:"Deep Learning"
3260:
3256:
3252:
3180:
3157:quantified-self
3129:Rainer Mühlhoff
3105:
3094:
3088:
3085:
3048:"Deep learning"
3042:
3040:
3030:
3018:
3007:
2935:
2902:
2842:
2836:
2828:
2776:
2756:backpropagation
2724:
2687:
2681:
2672:
2663:
2637:
2617:
2609:
2583:Google DeepMind
2579:
2571:fraud detection
2567:
2535:
2522:
2513:
2475:X to an output
2461:random variable
2449:
2411:
2405:
2396:
2390:
2367:
2361:
2313:
2302:
2293:
2226:
2220:
2185:
2142:
2140:
2138:Computer vision
2134:
1943:
1910:
1904:
1899:
1884:frequency combs
1835:cloud computing
1833:cellphones and
1806:
1757:
1753:
1751:
1748:
1747:
1726:
1722:
1720:
1717:
1716:
1698:
1646:
1615:backpropagation
1553:
1552:
1551:
1550:
1549:
1547:
1535:
1527:
1526:
1511:
1500:
1494:
1492:Neural networks
1478:Geoffrey Hinton
1343:Geoffrey Hinton
1335:Alex Krizhevsky
1290:GeForce GTX 280
1267:
1230:, Osindero and
1190:
1101:Geoffrey Hinton
1093:Terry Sejnowski
1047:Sepp Hochreiter
1027:a higher level
992:(1986) and the
948:
924:Henry J. Kelley
912:Backpropagation
885:introduced the
868:Shun'ichi Amari
819:Shun'ichi Amari
787:
782:
668:
666:Interpretations
549:neural networks
545:
522:climate science
494:computer vision
455:semi-supervised
434:neural networks
428:is a subset of
419:
390:
389:
380:
372:
371:
347:
337:
336:
308:Control problem
288:
278:
277:
189:
179:
178:
139:
131:
130:
101:Computer vision
76:
35:
28:
23:
22:
15:
12:
11:
5:
15059:
15049:
15048:
15031:
15030:
15028:
15027:
15026:
15025:
15020:
15007:
15006:
15005:
15000:
14986:
14983:
14982:
14980:
14979:
14974:
14969:
14964:
14959:
14954:
14949:
14944:
14939:
14934:
14929:
14924:
14919:
14914:
14909:
14903:
14901:
14897:
14896:
14894:
14893:
14888:
14883:
14878:
14873:
14868:
14863:
14858:
14853:
14847:
14845:
14841:
14840:
14838:
14837:
14835:Ilya Sutskever
14832:
14827:
14822:
14817:
14812:
14807:
14802:
14800:Demis Hassabis
14797:
14792:
14790:Ian Goodfellow
14787:
14782:
14776:
14774:
14770:
14769:
14766:
14765:
14763:
14762:
14757:
14756:
14755:
14745:
14740:
14735:
14730:
14725:
14720:
14715:
14709:
14707:
14703:
14702:
14700:
14699:
14694:
14689:
14684:
14679:
14674:
14669:
14664:
14659:
14654:
14649:
14644:
14639:
14634:
14629:
14624:
14619:
14618:
14617:
14607:
14602:
14597:
14592:
14587:
14581:
14579:
14575:
14574:
14572:
14571:
14566:
14565:
14564:
14559:
14549:
14548:
14547:
14542:
14537:
14527:
14522:
14517:
14512:
14507:
14502:
14497:
14492:
14487:
14481:
14479:
14472:
14468:
14467:
14465:
14464:
14459:
14454:
14449:
14444:
14439:
14434:
14428:
14426:
14422:
14421:
14419:
14418:
14413:
14408:
14403:
14398:
14392:
14390:
14386:
14385:
14383:
14382:
14381:
14380:
14373:Language model
14370:
14365:
14360:
14359:
14358:
14348:
14347:
14346:
14335:
14333:
14329:
14328:
14326:
14325:
14323:Autoregression
14320:
14315:
14314:
14313:
14303:
14301:Regularization
14298:
14297:
14296:
14291:
14286:
14276:
14271:
14266:
14264:Loss functions
14261:
14256:
14251:
14246:
14241:
14240:
14239:
14229:
14224:
14223:
14222:
14211:
14209:
14205:
14204:
14202:
14201:
14199:Inductive bias
14196:
14191:
14186:
14181:
14176:
14171:
14166:
14161:
14153:
14151:
14145:
14144:
14139:
14138:
14131:
14124:
14116:
14109:
14108:
14077:
14063:Bengio, Yoshua
14055:
14049:
14036:
14030:
14016:
14014:
14011:
14009:
14008:
13970:
13914:
13895:
13857:
13818:
13788:
13758:
13732:
13723:
13702:(4): 259–362.
13686:
13665:
13644:
13607:
13576:
13549:
13533:The New Yorker
13519:
13488:
13457:
13427:
13412:
13386:
13355:
13325:
13242:
13185:
13154:
13095:
13030:
12995:(3): 356–365.
12979:
12944:(4): 481–487.
12928:
12891:(2): 383–394.
12871:
12796:
12761:(9): 657–664.
12745:
12678:
12651:(5): 895–938.
12635:
12568:
12555:
12525:10.1.1.41.7854
12518:(4): 537–556.
12502:
12467:
12460:
12440:
12397:
12368:
12343:
12275:
12210:
12153:
12084:
12027:
11970:
11905:
11870:
11827:
11760:
11735:
11705:
11670:
11636:
11606:
11591:
11565:
11506:
11449:
11423:
11393:
11331:
11268:
11244:
11220:
11183:(2): 361–370.
11163:
11104:
11077:
11041:
11010:
10957:
10923:
10902:
10872:
10829:
10808:
10778:
10743:
10722:
10696:
10671:
10662:
10632:
10603:
10572:(5): 505–513.
10552:
10503:
10494:
10458:
10425:
10403:
10365:
10334:
10304:
10253:
10218:
10184:
10165:(3): 530–539.
10145:
10114:
10083:
10046:
10010:
9974:
9953:
9915:
9877:
9860:
9839:
9794:
9778:yann.lecun.com
9765:
9743:
9712:
9654:
9623:
9573:
9556:Robinson, Tony
9547:
9533:
9505:
9447:
9390:
9365:
9340:
9297:
9279:
9249:
9218:
9188:
9171:airesearch.com
9157:
9136:
9127:
9074:
9047:
9015:
8985:
8955:
8919:
8904:
8858:
8843:
8810:
8771:
8738:
8687:
8635:
8611:
8587:
8528:
8467:
8436:
8406:
8376:
8350:
8290:
8270:978-1492671206
8269:
8251:
8240:awards.acm.org
8227:
8191:
8170:
8145:
8130:
8094:
8065:
8053:
8017:
7991:
7970:
7944:
7909:
7888:
7881:
7839:
7814:
7793:
7769:
7747:
7724:
7702:
7670:
7649:
7628:
7613:
7587:
7553:
7509:
7448:
7441:
7411:
7387:
7357:
7319:
7288:
7242:
7212:
7179:
7149:
7142:
7117:
7079:
7027:
7005:
6956:
6909:
6870:
6845:
6776:
6709:
6692:
6672:
6666:10.1.1.75.6306
6640:
6601:
6571:
6562:
6551:(2): 181–192.
6532:
6521:(2): 225–254.
6505:
6494:(2): 331–339.
6478:
6452:
6401:
6352:(3): 328–339.
6329:
6311:
6286:
6267:(4): 899–916.
6251:
6224:(4): 865–884.
6204:
6197:
6179:
6117:
6073:(5): 889–904.
6042:
6035:
6004:
5971:(1): 147–169.
5951:
5892:
5873:(3): 230–247.
5854:
5849:Proc. SAB'1991
5833:
5826:
5800:
5769:
5762:
5740:
5710:
5689:
5670:(2): 234–242.
5644:
5630:(April 1991).
5614:
5595:(2): 179–211.
5575:
5552:
5523:10.1.1.32.9552
5493:
5446:
5419:(29): 4211–7.
5413:Applied Optics
5399:
5372:(32): 4790–7.
5366:Applied Optics
5352:
5335:
5316:
5298:
5278:
5258:
5211:
5204:
5186:
5146:
5113:
5104:
5085:(2): 146–160.
5066:
5048:
5043:10.2514/8.5282
5018:
5011:
4991:
4964:(4): 193–202.
4948:
4909:
4888:
4885:(16): 279–307.
4866:
4830:
4783:
4772:(2): 207–219.
4752:
4745:
4725:
4710:
4689:
4662:(6): 386–408.
4642:
4627:
4604:
4597:
4571:
4512:
4475:
4453:
4434:(4): 883–893.
4418:
4397:
4376:
4353:
4324:(2): 233–268.
4308:
4297:(4): 322–333.
4278:
4271:
4244:
4219:(3–4): 1–199.
4187:
4128:
4105:
4098:
4075:
4068:
4045:
4026:(2): 251–257.
4005:
3971:(4): 303–314.
3941:
3892:
3883:
3869:
3841:
3819:
3775:
3741:
3734:
3728:. IGI Global.
3714:
3646:
3584:
3510:
3485:
3455:
3416:
3401:
3362:
3301:
3274:(4): 357–363.
3253:
3251:
3248:
3247:
3246:
3241:
3236:
3231:
3226:
3221:
3216:
3211:
3206:
3201:
3196:
3191:
3186:
3179:
3176:
3107:
3106:
3021:
3019:
3012:
3006:
3003:
2999:data poisoning
2934:
2931:
2901:
2898:
2835:
2832:
2827:
2824:
2775:
2772:
2739:) support the
2723:
2720:
2708:ovarian cancer
2683:Main article:
2680:
2677:
2671:
2668:
2662:
2659:
2636:
2633:
2616:
2613:
2608:
2605:
2578:
2575:
2566:
2563:
2534:
2531:
2521:
2518:
2512:
2509:
2448:
2445:
2419:bioinformatics
2409:Bioinformatics
2407:Main article:
2404:
2403:Bioinformatics
2401:
2392:Main article:
2389:
2386:
2363:Main article:
2360:
2357:
2306:Drug discovery
2301:
2298:
2292:
2289:
2262:word embedding
2233:word embedding
2222:Main article:
2219:
2216:
2215:
2214:
2211:
2205:
2184:
2181:
2168:MNIST database
2136:Main article:
2133:
2130:
2090:
2089:
2086:
2080:
2070:
2060:
2057:
2054:
2051:
2042:
2041:
2038:
2034:
2033:
2030:
2026:
2025:
2022:
2018:
2017:
2014:
2010:
2009:
2006:
2002:
2001:
1998:
1994:
1993:
1990:
1986:
1985:
1982:
1978:
1977:
1974:
1970:
1969:
1966:
1962:
1961:
1958:
1954:
1953:
1950:
1946:
1945:
1940:
1906:Main article:
1903:
1900:
1898:
1895:
1854:semiconductors
1805:
1802:
1760:
1756:
1729:
1725:
1709:Regularization
1697:
1694:
1645:
1642:
1630:social network
1544:false positive
1536:
1529:
1528:
1512:
1505:
1504:
1503:
1502:
1501:
1496:Main article:
1493:
1490:
1406:Ian Goodfellow
1355:Karen Simonyan
1339:Ilya Sutskever
1266:
1263:
1259:decision trees
1189:
1186:
1076:that models a
990:Jordan network
947:
944:
928:control theory
904:introduced by
789:There are two
786:
783:
781:
778:
746:generalization
691:George Cybenko
667:
664:
544:
541:
510:bioinformatics
421:
420:
418:
417:
410:
403:
395:
392:
391:
388:
387:
381:
378:
377:
374:
373:
370:
369:
364:
359:
354:
348:
343:
342:
339:
338:
335:
334:
329:
324:
319:
314:
305:
300:
295:
289:
284:
283:
280:
279:
276:
275:
270:
265:
260:
255:
254:
253:
243:
238:
233:
232:
231:
226:
221:
211:
206:
204:Earth sciences
201:
196:
194:Bioinformatics
190:
185:
184:
181:
180:
177:
176:
171:
166:
161:
156:
151:
146:
140:
137:
136:
133:
132:
129:
128:
123:
118:
113:
108:
103:
98:
93:
88:
83:
77:
72:
71:
68:
67:
57:
56:
50:
49:
26:
9:
6:
4:
3:
2:
15058:
15047:
15046:Deep learning
15044:
15043:
15041:
15024:
15021:
15019:
15016:
15015:
15008:
15004:
15001:
14999:
14996:
14995:
14992:
14988:
14987:
14984:
14978:
14975:
14973:
14970:
14968:
14965:
14963:
14960:
14958:
14955:
14953:
14950:
14948:
14945:
14943:
14940:
14938:
14935:
14933:
14930:
14928:
14925:
14923:
14920:
14918:
14915:
14913:
14910:
14908:
14905:
14904:
14902:
14900:Architectures
14898:
14892:
14889:
14887:
14884:
14882:
14879:
14877:
14874:
14872:
14869:
14867:
14864:
14862:
14859:
14857:
14854:
14852:
14849:
14848:
14846:
14844:Organizations
14842:
14836:
14833:
14831:
14828:
14826:
14823:
14821:
14818:
14816:
14813:
14811:
14808:
14806:
14803:
14801:
14798:
14796:
14793:
14791:
14788:
14786:
14783:
14781:
14780:Yoshua Bengio
14778:
14777:
14775:
14771:
14761:
14760:Robot control
14758:
14754:
14751:
14750:
14749:
14746:
14744:
14741:
14739:
14736:
14734:
14731:
14729:
14726:
14724:
14721:
14719:
14716:
14714:
14711:
14710:
14708:
14704:
14698:
14695:
14693:
14690:
14688:
14685:
14683:
14680:
14678:
14677:Chinchilla AI
14675:
14673:
14670:
14668:
14665:
14663:
14660:
14658:
14655:
14653:
14650:
14648:
14645:
14643:
14640:
14638:
14635:
14633:
14630:
14628:
14625:
14623:
14620:
14616:
14613:
14612:
14611:
14608:
14606:
14603:
14601:
14598:
14596:
14593:
14591:
14588:
14586:
14583:
14582:
14580:
14576:
14570:
14567:
14563:
14560:
14558:
14555:
14554:
14553:
14550:
14546:
14543:
14541:
14538:
14536:
14533:
14532:
14531:
14528:
14526:
14523:
14521:
14518:
14516:
14513:
14511:
14508:
14506:
14503:
14501:
14498:
14496:
14493:
14491:
14488:
14486:
14483:
14482:
14480:
14476:
14473:
14469:
14463:
14460:
14458:
14455:
14453:
14450:
14448:
14445:
14443:
14440:
14438:
14435:
14433:
14430:
14429:
14427:
14423:
14417:
14414:
14412:
14409:
14407:
14404:
14402:
14399:
14397:
14394:
14393:
14391:
14387:
14379:
14376:
14375:
14374:
14371:
14369:
14366:
14364:
14361:
14357:
14356:Deep learning
14354:
14353:
14352:
14349:
14345:
14342:
14341:
14340:
14337:
14336:
14334:
14330:
14324:
14321:
14319:
14316:
14312:
14309:
14308:
14307:
14304:
14302:
14299:
14295:
14292:
14290:
14287:
14285:
14282:
14281:
14280:
14277:
14275:
14272:
14270:
14267:
14265:
14262:
14260:
14257:
14255:
14252:
14250:
14247:
14245:
14244:Hallucination
14242:
14238:
14235:
14234:
14233:
14230:
14228:
14225:
14221:
14218:
14217:
14216:
14213:
14212:
14210:
14206:
14200:
14197:
14195:
14192:
14190:
14187:
14185:
14182:
14180:
14177:
14175:
14172:
14170:
14167:
14165:
14162:
14160:
14159:
14155:
14154:
14152:
14150:
14146:
14137:
14132:
14130:
14125:
14123:
14118:
14117:
14114:
14105:
14099:
14084:
14080:
14074:
14071:. MIT Press.
14070:
14069:
14068:Deep Learning
14064:
14060:
14056:
14052:
14050:9780262048644
14046:
14042:
14037:
14033:
14027:
14023:
14018:
14017:
13996:
13992:
13988:
13984:
13980:
13974:
13966:
13962:
13958:
13954:
13949:
13944:
13940:
13936:
13932:
13925:
13923:
13921:
13919:
13910:
13906:
13899:
13884:
13880:
13876:
13872:
13868:
13861:
13845:
13841:
13837:
13831:
13829:
13827:
13825:
13823:
13806:
13802:
13801:The Daily Dot
13798:
13792:
13776:
13772:
13768:
13762:
13748:on 2017-12-30
13747:
13743:
13736:
13727:
13719:
13715:
13710:
13705:
13701:
13697:
13690:
13681:
13676:
13669:
13660:
13655:
13648:
13630:
13623:
13616:
13614:
13612:
13595:
13591:
13587:
13580:
13564:
13560:
13553:
13538:
13534:
13530:
13523:
13507:
13503:
13499:
13492:
13476:
13472:
13468:
13461:
13445:
13441:
13437:
13431:
13423:
13419:
13415:
13409:
13405:
13401:
13397:
13390:
13374:
13370:
13366:
13359:
13343:
13339:
13335:
13329:
13321:
13314:
13310:
13306:
13302:
13298:
13294:
13290:
13286:
13282:
13278:
13274:
13270:
13269:
13264:
13260:
13256:
13252:
13251:Silver, David
13246:
13238:
13234:
13230:
13226:
13221:
13216:
13212:
13208:
13204:
13200:
13196:
13189:
13173:
13169:
13165:
13158:
13150:
13146:
13141:
13136:
13132:
13128:
13123:
13118:
13114:
13110:
13106:
13099:
13091:
13087:
13083:
13079:
13074:
13069:
13065:
13061:
13057:
13053:
13049:
13045:
13041:
13034:
13026:
13022:
13018:
13014:
13010:
13006:
13002:
12998:
12994:
12990:
12983:
12975:
12971:
12967:
12963:
12959:
12955:
12951:
12947:
12943:
12939:
12932:
12924:
12920:
12916:
12912:
12908:
12904:
12899:
12894:
12890:
12886:
12882:
12875:
12867:
12863:
12859:
12855:
12850:
12845:
12841:
12837:
12832:
12827:
12823:
12819:
12815:
12811:
12807:
12800:
12792:
12788:
12784:
12780:
12776:
12772:
12768:
12764:
12760:
12756:
12749:
12741:
12737:
12733:
12729:
12724:
12719:
12715:
12711:
12706:
12701:
12697:
12693:
12689:
12682:
12674:
12670:
12666:
12662:
12658:
12654:
12650:
12646:
12639:
12631:
12627:
12622:
12617:
12613:
12609:
12604:
12599:
12595:
12591:
12587:
12583:
12579:
12572:
12565:
12559:
12551:
12547:
12543:
12539:
12535:
12531:
12526:
12521:
12517:
12513:
12506:
12498:
12494:
12490:
12486:
12482:
12478:
12471:
12463:
12457:
12454:. MIT Press.
12453:
12452:
12444:
12436:
12432:
12428:
12424:
12420:
12416:
12412:
12408:
12401:
12392:
12387:
12383:
12379:
12372:
12358:
12354:
12347:
12339:
12335:
12331:
12327:
12322:
12317:
12313:
12309:
12304:
12299:
12295:
12291:
12287:
12279:
12271:
12267:
12263:
12259:
12255:
12251:
12247:
12243:
12238:
12233:
12229:
12225:
12221:
12214:
12206:
12202:
12198:
12194:
12190:
12186:
12181:
12176:
12172:
12168:
12164:
12157:
12149:
12145:
12140:
12135:
12130:
12125:
12121:
12117:
12112:
12107:
12103:
12099:
12095:
12088:
12080:
12076:
12071:
12066:
12062:
12058:
12054:
12050:
12046:
12042:
12038:
12031:
12023:
12019:
12015:
12011:
12006:
12001:
11997:
11993:
11989:
11985:
11981:
11974:
11966:
11962:
11958:
11954:
11950:
11946:
11941:
11936:
11932:
11928:
11924:
11920:
11916:
11909:
11893:
11889:
11885:
11879:
11877:
11875:
11866:
11862:
11858:
11854:
11850:
11846:
11842:
11838:
11831:
11823:
11819:
11814:
11809:
11805:
11801:
11796:
11791:
11787:
11783:
11779:
11775:
11771:
11764:
11750:
11746:
11739:
11724:
11720:
11716:
11709:
11701:
11697:
11693:
11689:
11685:
11681:
11674:
11656:
11649:
11648:
11640:
11624:
11620:
11619:FloydHub Blog
11616:
11610:
11602:
11598:
11594:
11588:
11584:
11580:
11576:
11569:
11561:
11557:
11552:
11547:
11543:
11539:
11534:
11529:
11525:
11521:
11517:
11510:
11502:
11498:
11494:
11490:
11485:
11480:
11476:
11472:
11468:
11464:
11460:
11453:
11438:
11434:
11430:
11426:
11424:9781538610343
11420:
11416:
11412:
11408:
11404:
11397:
11389:
11385:
11381:
11377:
11373:
11369:
11365:
11361:
11356:
11351:
11347:
11343:
11335:
11327:
11323:
11319:
11315:
11311:
11307:
11302:
11297:
11293:
11289:
11282:
11275:
11273:
11258:
11255:
11248:
11234:
11230:
11224:
11216:
11212:
11207:
11202:
11198:
11194:
11190:
11186:
11182:
11178:
11174:
11167:
11159:
11155:
11151:
11147:
11142:
11137:
11132:
11127:
11123:
11119:
11115:
11108:
11092:
11088:
11084:
11080:
11078:9781450328944
11074:
11070:
11066:
11062:
11058:
11054:
11053:
11045:
11029:
11025:
11021:
11014:
11006:
11002:
10997:
10992:
10987:
10986:10.2196/12957
10982:
10979:(5): e12957.
10978:
10974:
10973:
10968:
10961:
10943:
10936:
10935:
10927:
10918:
10913:
10906:
10891:
10887:
10883:
10876:
10868:
10864:
10860:
10856:
10852:
10848:
10844:
10840:
10833:
10824:
10819:
10812:
10796:
10792:
10788:
10782:
10766:
10762:
10761:
10756:
10750:
10748:
10738:
10733:
10726:
10710:
10706:
10700:
10685:
10681:
10675:
10666:
10650:
10646:
10642:
10636:
10621:
10617:
10613:
10607:
10599:
10595:
10590:
10585:
10580:
10575:
10571:
10567:
10563:
10556:
10548:
10544:
10540:
10536:
10531:
10526:
10522:
10518:
10514:
10507:
10498:
10479:
10472:
10465:
10463:
10446:
10442:
10441:
10436:
10429:
10420:
10415:
10407:
10391:
10387:
10383:
10376:
10374:
10372:
10370:
10353:
10349:
10345:
10338:
10322:
10318:
10314:
10308:
10300:
10296:
10291:
10286:
10281:
10276:
10272:
10268:
10264:
10257:
10249:
10245:
10241:
10237:
10234:(12): e3259.
10233:
10229:
10222:
10206:
10202:
10198:
10191:
10189:
10180:
10176:
10172:
10168:
10164:
10160:
10156:
10149:
10133:
10129:
10125:
10118:
10102:
10098:
10094:
10087:
10068:
10064:
10057:
10050:
10032:
10028:
10021:
10014:
9995:
9988:
9981:
9979:
9969:
9964:
9957:
9948:
9943:
9939:
9935:
9931:
9924:
9922:
9920:
9910:
9905:
9901:
9897:
9893:
9886:
9884:
9882:
9875:
9874:
9869:
9864:
9855:
9850:
9843:
9835:
9831:
9827:
9823:
9818:
9813:
9809:
9805:
9798:
9783:
9779:
9775:
9769:
9760:
9755:
9747:
9731:
9727:
9723:
9716:
9698:
9694:
9690:
9685:
9680:
9676:
9672:
9665:
9658:
9650:
9646:
9642:
9638:
9635:: 1915–1919.
9634:
9627:
9612:
9608:
9604:
9600:
9596:
9592:
9588:
9584:
9577:
9569:
9565:
9561:
9557:
9551:
9536:
9534:1-58563-019-5
9530:
9526:
9522:
9518:
9517:
9509:
9501:
9497:
9493:
9489:
9485:
9481:
9476:
9471:
9467:
9463:
9456:
9454:
9452:
9443:
9439:
9434:
9429:
9425:
9421:
9417:
9413:
9409:
9405:
9401:
9394:
9379:
9375:
9369:
9355:
9351:
9344:
9335:
9330:
9325:
9320:
9316:
9312:
9308:
9301:
9293:
9289:
9283:
9267:
9263:
9259:
9253:
9237:
9233:
9229:
9222:
9206:
9202:
9198:
9192:
9176:
9172:
9168:
9161:
9154:
9150:
9146:
9140:
9131:
9123:
9119:
9115:
9111:
9107:
9103:
9098:
9093:
9089:
9085:
9078:
9062:
9058:
9054:
9050:
9048:9781450351140
9044:
9040:
9036:
9032:
9031:
9026:
9019:
9004:
9000:
8996:
8989:
8973:
8969:
8965:
8959:
8941:
8937:
8930:
8923:
8915:
8911:
8907:
8901:
8897:
8893:
8888:
8883:
8878:
8873:
8869:
8862:
8854:
8850:
8846:
8840:
8836:
8832:
8828:
8824:
8820:
8814:
8806:
8802:
8798:
8794:
8790:
8786:
8782:
8775:
8759:
8755:
8751:
8745:
8743:
8734:
8730:
8726:
8722:
8718:
8714:
8710:
8706:
8702:
8698:
8691:
8673:
8669:
8665:
8661:
8657:
8654:: 1045–1048.
8653:
8646:
8639:
8630:
8625:
8618:
8616:
8606:
8601:
8594:
8592:
8573:
8569:
8565:
8560:
8555:
8551:
8544:
8537:
8535:
8533:
8517:
8513:
8509:
8505:
8501:
8497:
8493:
8489:
8485:
8481:
8474:
8472:
8455:
8451:
8447:
8440:
8425:
8422:. ICLR 2018.
8421:
8417:
8410:
8395:
8392:: 2553–2561.
8391:
8387:
8380:
8366:
8362:
8361:
8354:
8346:
8342:
8338:
8334:
8330:
8326:
8322:
8318:
8314:
8310:
8306:
8302:
8294:
8286:
8280:
8272:
8266:
8262:
8255:
8241:
8237:
8231:
8213:
8209:
8202:
8195:
8186:
8181:
8174:
8163:
8156:
8149:
8141:
8137:
8133:
8127:
8123:
8119:
8114:
8109:
8105:
8098:
8083:
8079:
8072:
8070:
8063:
8057:
8048:
8043:
8039:
8035:
8028:
8021:
8005:
8001:
7995:
7986:
7981:
7974:
7958:
7954:
7948:
7929:
7922:
7921:
7913:
7904:
7899:
7892:
7884:
7878:
7874:
7870:
7865:
7860:
7856:
7852:
7851:
7843:
7834:
7829:
7825:
7818:
7809:
7804:
7797:
7789:
7784:
7780:
7773:
7763:
7758:
7751:
7741:
7736:
7728:
7718:
7713:
7706:
7697:
7692:
7688:
7681:
7674:
7665:
7660:
7653:
7644:
7639:
7632:
7624:
7620:
7616:
7610:
7606:
7602:
7598:
7591:
7573:
7566:
7565:
7557:
7539:
7535:
7531:
7527:
7520:
7513:
7505:
7501:
7497:
7493:
7489:
7485:
7481:
7477:
7472:
7467:
7463:
7459:
7452:
7444:
7438:
7434:
7430:
7426:
7422:
7415:
7406:
7401:
7397:
7396:Sze, Vivienne
7391:
7377:
7373:
7372:
7364:
7362:
7353:
7349:
7345:
7341:
7337:
7333:
7326:
7324:
7307:
7303:
7299:
7292:
7277:
7273:
7269:
7265:
7261:
7257:
7253:
7246:
7231:
7227:
7223:
7216:
7201:
7197:
7193:
7186:
7184:
7167:
7163:
7159:
7153:
7145:
7139:
7135:
7134:
7126:
7124:
7122:
7102:
7099:. Microsoft.
7095:
7088:
7086:
7084:
7075:
7071:
7067:
7063:
7059:
7055:
7051:
7047:
7043:
7036:
7034:
7032:
7025:
7021:
7018:
7014:
7009:
7001:
6997:
6992:
6987:
6983:
6979:
6975:
6971:
6967:
6960:
6952:
6948:
6944:
6940:
6936:
6932:
6928:
6924:
6920:
6913:
6905:
6901:
6897:
6893:
6889:
6885:
6881:
6874:
6867:
6863:
6859:
6855:
6849:
6831:
6827:
6823:
6819:
6815:
6811:
6807:
6803:
6799:
6798:
6790:
6786:
6785:Hinton, G. E.
6780:
6764:
6760:
6756:
6752:
6748:
6744:
6740:
6736:
6732:
6728:
6724:
6720:
6713:
6706:
6702:
6696:
6689:
6685:
6682:
6676:
6667:
6662:
6658:
6654:
6650:
6644:
6626:
6622:
6615:
6608:
6606:
6589:
6585:
6581:
6575:
6566:
6558:
6554:
6550:
6546:
6539:
6537:
6528:
6524:
6520:
6516:
6509:
6501:
6497:
6493:
6489:
6482:
6467:
6463:
6456:
6448:
6444:
6440:
6436:
6432:
6428:
6424:
6420:
6416:
6412:
6405:
6387:
6383:
6379:
6375:
6371:
6367:
6363:
6359:
6355:
6351:
6347:
6340:
6333:
6318:
6314:
6312:9780780305328
6308:
6304:
6300:
6296:
6290:
6282:
6278:
6274:
6270:
6266:
6262:
6255:
6247:
6243:
6239:
6235:
6231:
6227:
6223:
6219:
6215:
6208:
6200:
6194:
6190:
6183:
6175:
6171:
6167:
6163:
6159:
6155:
6151:
6147:
6143:
6139:
6135:
6131:
6127:
6121:
6113:
6106:
6102:
6098:
6094:
6089:
6084:
6080:
6076:
6072:
6068:
6064:
6060:
6056:
6052:
6046:
6038:
6036:0-262-68053-X
6032:
6028:
6024:
6023:
6015:
6008:
6000:
5994:
5986:
5982:
5978:
5974:
5970:
5966:
5962:
5955:
5947:
5943:
5939:
5935:
5931:
5927:
5922:
5917:
5913:
5909:
5905:
5899:
5897:
5888:
5884:
5880:
5876:
5872:
5868:
5864:
5858:
5850:
5846:
5840:
5838:
5829:
5827:0-85296-721-7
5823:
5819:
5815:
5811:
5804:
5797:
5793:
5789:
5788:
5783:
5779:
5773:
5765:
5759:
5755:
5751:
5744:
5737:
5733:
5729:
5725:
5719:
5717:
5715:
5702:
5701:
5693:
5685:
5681:
5677:
5673:
5669:
5665:
5658:
5651:
5649:
5640:
5633:
5629:
5623:
5621:
5619:
5610:
5606:
5602:
5598:
5594:
5590:
5586:
5579:
5571:
5567:
5563:
5556:
5541:
5537:
5533:
5529:
5524:
5519:
5515:
5511:
5504:
5497:
5489:
5485:
5481:
5477:
5473:
5469:
5466:(4): 517–24.
5465:
5461:
5457:
5450:
5442:
5438:
5434:
5430:
5426:
5422:
5418:
5414:
5410:
5403:
5395:
5391:
5387:
5383:
5379:
5375:
5371:
5367:
5363:
5356:
5349:
5345:
5339:
5331:
5327:
5320:
5313:
5312:
5307:
5302:
5291:
5290:
5282:
5275:
5271:
5268:
5262:
5254:
5250:
5246:
5242:
5238:
5234:
5230:
5226:
5222:
5215:
5207:
5205:0-471-59897-6
5201:
5197:
5190:
5171:
5167:
5160:
5156:
5150:
5134:
5130:
5126:
5120:
5118:
5108:
5100:
5096:
5092:
5088:
5084:
5080:
5076:
5070:
5062:
5058:
5052:
5044:
5040:
5036:
5032:
5028:
5022:
5014:
5012:9780598818461
5008:
5004:
5003:
4995:
4987:
4983:
4979:
4975:
4971:
4967:
4963:
4959:
4952:
4944:
4940:
4936:
4932:
4928:
4924:
4920:
4913:
4904:
4899:
4892:
4884:
4880:
4876:
4870:
4861:
4856:
4852:
4848:
4844:
4840:
4834:
4816:
4812:
4808:
4804:
4797:
4790:
4788:
4779:
4775:
4771:
4767:
4763:
4756:
4748:
4742:
4738:
4737:
4729:
4721:
4714:
4706:
4702:
4696:
4694:
4685:
4681:
4677:
4673:
4669:
4665:
4661:
4657:
4653:
4646:
4638:
4631:
4623:
4619:
4615:
4608:
4600:
4594:
4590:
4586:
4582:
4575:
4567:
4563:
4558:
4553:
4548:
4543:
4539:
4535:
4531:
4527:
4523:
4516:
4507:
4502:
4498:
4492:
4490:
4488:
4486:
4484:
4482:
4480:
4471:
4467:
4460:
4458:
4449:
4445:
4441:
4437:
4433:
4429:
4422:
4414:
4410:
4404:
4402:
4383:
4379:
4373:
4366:
4365:
4357:
4349:
4345:
4341:
4337:
4332:
4327:
4323:
4319:
4312:
4304:
4300:
4296:
4292:
4285:
4283:
4274:
4268:
4265:. MIT Press.
4264:
4263:
4255:
4253:
4251:
4249:
4230:
4226:
4222:
4218:
4214:
4207:
4200:
4198:
4196:
4194:
4192:
4183:
4179:
4174:
4169:
4164:
4159:
4155:
4151:
4147:
4143:
4139:
4132:
4125:
4121:
4118:
4112:
4110:
4101:
4095:
4091:
4090:
4082:
4080:
4071:
4065:
4061:
4060:
4052:
4050:
4041:
4037:
4033:
4029:
4025:
4021:
4014:
4012:
4010:
3998:
3994:
3990:
3986:
3982:
3978:
3974:
3970:
3966:
3965:
3957:
3950:
3948:
3946:
3937:
3933:
3929:
3925:
3920:
3915:
3911:
3907:
3903:
3896:
3887:
3872:
3866:
3862:
3858:
3854:
3853:
3845:
3839:
3835:
3832:
3828:
3823:
3814:
3809:
3805:
3801:
3797:
3793:
3789:
3782:
3780:
3761:
3754:
3753:
3745:
3737:
3731:
3727:
3726:
3718:
3710:
3706:
3702:
3698:
3694:
3690:
3685:
3680:
3676:
3672:
3665:
3663:
3661:
3659:
3657:
3655:
3653:
3651:
3642:
3638:
3634:
3630:
3626:
3622:
3617:
3612:
3608:
3604:
3597:
3595:
3593:
3591:
3589:
3569:
3565:
3561:
3557:
3553:
3548:
3543:
3539:
3535:
3528:
3521:
3519:
3517:
3515:
3499:
3495:
3489:
3473:
3469:
3465:
3459:
3441:
3437:
3430:
3423:
3421:
3412:
3408:
3404:
3398:
3394:
3390:
3385:
3380:
3376:
3369:
3367:
3358:
3354:
3350:
3346:
3342:
3338:
3334:
3330:
3326:
3322:
3315:
3308:
3306:
3297:
3293:
3289:
3285:
3281:
3277:
3273:
3269:
3265:
3258:
3254:
3245:
3242:
3240:
3237:
3235:
3234:Sparse coding
3232:
3230:
3227:
3225:
3222:
3220:
3217:
3215:
3212:
3210:
3207:
3205:
3202:
3200:
3197:
3195:
3192:
3190:
3187:
3185:
3182:
3181:
3175:
3173:
3168:
3166:
3162:
3158:
3154:
3150:
3146:
3145:tagging faces
3142:
3138:
3134:
3130:
3126:
3122:
3118:
3114:
3103:
3100:
3092:
3081:
3078:
3074:
3071:
3067:
3064:
3060:
3057:
3053:
3050: –
3049:
3045:
3044:Find sources:
3038:
3034:
3028:
3027:
3022:This section
3020:
3016:
3011:
3010:
3002:
3000:
2995:
2993:
2988:
2986:
2982:
2978:
2974:
2969:
2967:
2963:
2959:
2954:
2953:
2949:
2943:
2941:
2930:
2928:
2924:
2920:
2916:
2912:
2908:
2897:
2895:
2893:
2885:
2883:
2879:
2875:
2871:
2867:
2861:
2859:
2855:
2850:
2848:
2841:
2831:
2823:
2821:
2817:
2812:
2809:
2807:
2803:
2799:
2795:
2791:
2786:
2784:
2780:
2771:
2767:
2765:
2761:
2757:
2752:
2750:
2746:
2742:
2738:
2733:
2729:
2719:
2717:
2713:
2709:
2705:
2701:
2697:
2692:
2686:
2676:
2667:
2658:
2656:
2651:
2649:
2645:
2641:
2632:
2630:
2626:
2622:
2612:
2604:
2601:
2597:
2592:
2588:
2584:
2574:
2572:
2562:
2560:
2556:
2552:
2548:
2544:
2540:
2530:
2527:
2517:
2508:
2506:
2502:
2498:
2494:
2490:
2486:
2482:
2478:
2474:
2470:
2466:
2462:
2458:
2454:
2444:
2442:
2438:
2433:
2431:
2426:
2424:
2423:gene ontology
2421:, to predict
2420:
2416:
2410:
2400:
2395:
2385:
2383:
2379:
2375:
2371:
2366:
2356:
2353:
2348:
2346:
2342:
2338:
2333:
2331:
2330:toxic effects
2327:
2323:
2319:
2318:toxic effects
2311:
2307:
2297:
2288:
2285:
2281:
2277:
2273:
2269:
2267:
2263:
2258:
2256:
2252:
2248:
2244:
2240:
2239:
2234:
2229:
2225:
2212:
2209:
2206:
2203:
2202:
2201:
2194:
2189:
2180:
2176:
2172:
2169:
2162:
2158:
2139:
2129:
2127:
2123:
2119:
2115:
2111:
2107:
2103:
2099:
2095:
2087:
2084:
2081:
2078:
2074:
2071:
2068:
2064:
2061:
2058:
2055:
2052:
2049:
2048:
2047:
2039:
2036:
2035:
2031:
2028:
2027:
2023:
2020:
2019:
2015:
2012:
2011:
2007:
2004:
2003:
1999:
1996:
1995:
1991:
1988:
1987:
1983:
1980:
1979:
1975:
1972:
1971:
1967:
1964:
1963:
1959:
1956:
1955:
1951:
1948:
1947:
1942:Percent phone
1941:
1938:
1937:
1934:
1931:
1927:
1923:
1919:
1914:
1909:
1894:
1892:
1889:
1885:
1881:
1877:
1873:
1870:
1865:
1863:
1860:
1859:floating-gate
1855:
1850:
1848:
1844:
1841:(TPU) in the
1840:
1836:
1832:
1828:
1824:
1819:
1816:
1811:
1801:
1799:
1794:
1792:
1788:
1784:
1783:learning rate
1779:
1776:
1758:
1754:
1745:
1727:
1723:
1714:
1710:
1705:
1703:
1693:
1691:
1687:
1683:
1681:
1677:
1673:
1669:
1665:
1663:
1659:
1654:
1650:
1641:
1637:
1635:
1631:
1627:
1623:
1618:
1616:
1610:
1606:
1604:
1600:
1596:
1592:
1588:
1583:
1581:
1577:
1572:
1568:
1566:
1565:connectionist
1561:
1557:
1545:
1541:
1533:
1524:
1520:
1516:
1509:
1499:
1489:
1487:
1483:
1479:
1475:
1474:Yoshua Bengio
1471:
1469:
1465:
1461:
1457:
1453:
1448:
1446:
1442:
1437:
1435:
1431:
1427:
1423:
1419:
1415:
1411:
1407:
1403:
1399:
1397:
1393:
1389:
1384:
1382:
1378:
1373:
1371:
1366:
1364:
1361:and Google's
1360:
1356:
1352:
1348:
1344:
1340:
1336:
1332:
1327:
1325:
1321:
1317:
1312:
1310:
1306:
1302:
1298:
1293:
1291:
1287:
1282:
1278:
1271:
1262:
1260:
1256:
1250:
1246:
1244:
1240:
1236:
1233:
1229:
1225:
1220:
1218:
1214:
1210:
1206:
1201:
1199:
1195:
1194:Gabor filters
1185:
1183:
1179:
1175:
1170:
1168:
1164:
1160:
1156:
1152:
1148:
1144:
1139:
1137:
1133:
1132:mixture model
1129:
1124:
1122:
1118:
1114:
1110:
1106:
1102:
1098:
1094:
1089:
1087:
1083:
1079:
1075:
1071:
1070:zero-sum game
1067:
1062:
1060:
1056:
1052:
1048:
1044:
1042:
1038:
1034:
1030:
1026:
1022:
1018:
1014:
1010:
1006:
1001:
999:
995:
994:Elman network
991:
987:
983:
981:
977:
973:
969:
965:
961:
957:
953:
943:
941:
937:
933:
929:
925:
921:
917:
913:
909:
907:
903:
899:
894:
892:
888:
884:
879:
877:
873:
869:
865:
861:
856:
853:
849:
844:
842:
838:
834:
832:
828:
824:
823:John Hopfield
820:
816:
812:
808:
804:
800:
796:
792:
777:
775:
771:
767:
763:
759:
755:
751:
747:
743:
739:
735:
731:
727:
726:probabilistic
722:
720:
716:
712:
707:
705:
701:
696:
692:
688:
684:
679:
677:
673:
663:
661:
657:
653:
652:Deep Learning
648:
646:
640:
638:
633:
631:
627:
623:
618:
615:
611:
607:
602:
600:
596:
592:
588:
584:
581:
576:
574:
570:
566:
562:
558:
554:
550:
540:
538:
534:
529:
527:
523:
519:
515:
511:
507:
503:
499:
495:
491:
487:
483:
479:
475:
471:
467:
462:
460:
456:
452:
447:
443:
439:
435:
431:
427:
426:Deep learning
416:
411:
409:
404:
402:
397:
396:
394:
393:
386:
383:
382:
376:
375:
368:
365:
363:
360:
358:
355:
353:
350:
349:
346:
341:
340:
333:
330:
328:
325:
323:
320:
318:
315:
313:
309:
306:
304:
301:
299:
296:
294:
291:
290:
287:
282:
281:
274:
271:
269:
266:
264:
261:
259:
256:
252:
251:Mental health
249:
248:
247:
244:
242:
239:
237:
234:
230:
227:
225:
222:
220:
217:
216:
215:
214:Generative AI
212:
210:
207:
205:
202:
200:
197:
195:
192:
191:
188:
183:
182:
175:
172:
170:
167:
165:
162:
160:
157:
155:
154:Deep learning
152:
150:
147:
145:
142:
141:
135:
134:
127:
124:
122:
119:
117:
114:
112:
109:
107:
104:
102:
99:
97:
94:
92:
89:
87:
84:
82:
79:
78:
75:
70:
69:
63:
59:
58:
55:
52:
51:
47:
46:
39:
33:
19:
14866:Hugging Face
14830:David Silver
14478:Audio–visual
14355:
14332:Applications
14311:Augmentation
14156:
14087:. Retrieved
14067:
14040:
14024:. Springer.
14021:
13999:. Retrieved
13982:
13973:
13938:
13934:
13908:
13898:
13887:. Retrieved
13870:
13860:
13848:. Retrieved
13839:
13809:. Retrieved
13800:
13791:
13779:. Retrieved
13770:
13761:
13750:. Retrieved
13746:the original
13735:
13726:
13699:
13695:
13689:
13668:
13647:
13636:. Retrieved
13598:. Retrieved
13590:The Guardian
13589:
13579:
13567:. Retrieved
13552:
13541:. Retrieved
13532:
13522:
13510:. Retrieved
13501:
13491:
13479:. Retrieved
13470:
13460:
13448:. Retrieved
13439:
13430:
13395:
13389:
13377:. Retrieved
13368:
13358:
13346:. Retrieved
13342:the original
13337:
13328:
13272:
13266:
13245:
13202:
13198:
13188:
13176:. Retrieved
13167:
13157:
13112:
13108:
13098:
13047:
13043:
13033:
12992:
12988:
12982:
12941:
12937:
12931:
12888:
12884:
12874:
12813:
12809:
12799:
12758:
12754:
12748:
12695:
12691:
12681:
12648:
12644:
12638:
12585:
12581:
12571:
12563:
12558:
12515:
12511:
12505:
12480:
12476:
12470:
12450:
12443:
12410:
12406:
12400:
12381:
12371:
12360:. Retrieved
12356:
12346:
12293:
12289:
12278:
12227:
12223:
12213:
12170:
12166:
12156:
12101:
12097:
12087:
12044:
12040:
12030:
11987:
11983:
11973:
11922:
11918:
11908:
11896:. Retrieved
11887:
11840:
11830:
11777:
11773:
11763:
11752:. Retrieved
11748:
11738:
11727:. Retrieved
11718:
11708:
11683:
11679:
11673:
11662:. Retrieved
11646:
11639:
11627:. Retrieved
11618:
11609:
11574:
11568:
11523:
11519:
11509:
11466:
11462:
11452:
11441:. Retrieved
11406:
11396:
11345:
11341:
11334:
11291:
11287:
11260:. Retrieved
11256:
11247:
11236:. Retrieved
11232:
11223:
11180:
11176:
11166:
11121:
11117:
11107:
11095:. Retrieved
11069:11311/964622
11051:
11044:
11032:. Retrieved
11023:
11013:
10976:
10970:
10960:
10949:. Retrieved
10933:
10926:
10905:
10894:. Retrieved
10885:
10875:
10842:
10838:
10832:
10811:
10799:. Retrieved
10790:
10781:
10769:. Retrieved
10758:
10725:
10713:. Retrieved
10709:the original
10699:
10688:. Retrieved
10674:
10665:
10653:. Retrieved
10644:
10635:
10624:. Retrieved
10615:
10606:
10569:
10565:
10555:
10520:
10516:
10506:
10497:
10485:. Retrieved
10478:the original
10449:. Retrieved
10438:
10428:
10406:
10394:. Retrieved
10385:
10356:. Retrieved
10347:
10337:
10325:. Retrieved
10316:
10307:
10270:
10266:
10256:
10231:
10227:
10221:
10209:. Retrieved
10200:
10162:
10158:
10148:
10136:. Retrieved
10127:
10117:
10105:. Retrieved
10096:
10086:
10074:. Retrieved
10062:
10049:
10038:. Retrieved
10026:
10013:
10001:. Retrieved
9956:
9937:
9933:
9899:
9895:
9871:
9863:
9842:
9807:
9803:
9797:
9786:. Retrieved
9777:
9768:
9746:
9734:. Retrieved
9725:
9715:
9704:. Retrieved
9674:
9670:
9657:
9632:
9626:
9615:. Retrieved
9590:
9586:
9576:
9559:
9550:
9538:. Retrieved
9515:
9508:
9468:(2): 52–58.
9465:
9461:
9410:(2): 72–77.
9407:
9403:
9393:
9382:. Retrieved
9380:. 2021-04-20
9377:
9368:
9357:. Retrieved
9353:
9343:
9314:
9310:
9300:
9291:
9282:
9270:. Retrieved
9261:
9252:
9240:. Retrieved
9231:
9221:
9209:. Retrieved
9200:
9191:
9179:. Retrieved
9170:
9160:
9139:
9130:
9087:
9083:
9077:
9065:. Retrieved
9029:
9018:
9007:. Retrieved
8998:
8988:
8976:. Retrieved
8967:
8958:
8947:. Retrieved
8935:
8922:
8867:
8861:
8826:
8813:
8788:
8784:
8774:
8762:. Retrieved
8754:ResearchGate
8753:
8700:
8696:
8690:
8679:. Retrieved
8651:
8638:
8579:. Retrieved
8549:
8520:. Retrieved
8487:
8483:
8458:. Retrieved
8454:the original
8449:
8439:
8428:. Retrieved
8419:
8409:
8398:. Retrieved
8389:
8379:
8369:, retrieved
8359:
8353:
8304:
8300:
8293:
8260:
8254:
8243:. Retrieved
8239:
8230:
8219:. Retrieved
8207:
8194:
8173:
8162:the original
8148:
8103:
8097:
8086:. Retrieved
8056:
8037:
8033:
8020:
8008:. Retrieved
7994:
7973:
7961:. Retrieved
7956:
7947:
7935:. Retrieved
7919:
7912:
7891:
7854:
7849:
7842:
7823:
7817:
7796:
7778:
7772:
7750:
7727:
7705:
7686:
7673:
7652:
7631:
7596:
7590:
7579:. Retrieved
7563:
7556:
7545:. Retrieved
7525:
7512:
7461:
7457:
7451:
7424:
7414:
7390:
7380:, retrieved
7370:
7335:
7331:
7310:. Retrieved
7301:
7291:
7280:. Retrieved
7255:
7245:
7234:. Retrieved
7225:
7215:
7204:. Retrieved
7195:
7170:. Retrieved
7161:
7152:
7136:. Springer.
7132:
7108:. Retrieved
7052:(6): 82–97.
7049:
7045:
7008:
6973:
6970:Scholarpedia
6969:
6959:
6926:
6922:
6912:
6887:
6883:
6873:
6865:
6848:
6837:. Retrieved
6801:
6795:
6779:
6767:. Retrieved
6726:
6722:
6712:
6704:
6700:
6695:
6675:
6656:
6649:Graves, Alex
6643:
6632:. Retrieved
6620:
6592:. Retrieved
6584:ResearchGate
6583:
6574:
6565:
6548:
6544:
6518:
6514:
6508:
6491:
6487:
6481:
6470:. Retrieved
6455:
6439:1721.1/51891
6417:(3): 75–80.
6414:
6410:
6404:
6393:. Retrieved
6349:
6345:
6332:
6321:. Retrieved
6302:
6295:Robinson, T.
6289:
6264:
6260:
6254:
6221:
6217:
6207:
6188:
6182:
6141:
6137:
6130:Dayan, Peter
6120:
6070:
6066:
6051:Peter, Dayan
6045:
6021:
6007:
5993:cite journal
5968:
5964:
5954:
5911:
5907:
5870:
5866:
5857:
5848:
5809:
5803:
5786:
5772:
5753:
5743:
5735:
5699:
5692:
5667:
5663:
5638:
5592:
5588:
5578:
5569:
5565:
5555:
5543:. Retrieved
5513:
5509:
5496:
5463:
5459:
5449:
5416:
5412:
5402:
5369:
5365:
5355:
5347:
5343:
5338:
5329:
5319:
5309:
5301:
5288:
5281:
5261:
5228:
5224:
5214:
5195:
5189:
5177:. Retrieved
5165:
5155:Werbos, Paul
5149:
5137:. Retrieved
5133:the original
5107:
5082:
5078:
5069:
5060:
5051:
5034:
5030:
5021:
5001:
4994:
4961:
4958:Biol. Cybern
4957:
4951:
4918:
4912:
4891:
4882:
4878:
4869:
4850:
4846:
4833:
4822:. Retrieved
4802:
4769:
4765:
4755:
4735:
4728:
4719:
4713:
4704:
4659:
4655:
4645:
4636:
4630:
4613:
4607:
4580:
4574:
4529:
4525:
4515:
4469:
4465:
4431:
4427:
4421:
4412:
4389:. Retrieved
4370:. Springer.
4363:
4356:
4321:
4317:
4311:
4294:
4290:
4261:
4236:. Retrieved
4216:
4212:
4145:
4141:
4131:
4088:
4058:
4023:
4019:
3997:the original
3968:
3962:
3909:
3905:
3895:
3886:
3874:. Retrieved
3851:
3844:
3827:Rina Dechter
3822:
3795:
3792:Scholarpedia
3791:
3767:. Retrieved
3751:
3744:
3724:
3717:
3674:
3670:
3606:
3602:
3575:. Retrieved
3568:the original
3540:(1): 1–127.
3537:
3533:
3502:. Retrieved
3500:. 2022-11-02
3497:
3488:
3476:. Retrieved
3467:
3458:
3447:. Retrieved
3435:
3374:
3324:
3320:
3271:
3267:
3257:
3169:
3133:gamification
3110:
3095:
3086:
3076:
3069:
3062:
3055:
3043:
3031:Please help
3026:verification
3023:
2996:
2989:
2970:
2955:
2951:
2944:
2936:
2933:Cyber threat
2903:
2892:The Guardian
2890:
2887:
2863:
2851:
2843:
2829:
2813:
2810:
2787:
2777:
2768:
2753:
2725:
2688:
2673:
2664:
2652:
2638:
2618:
2610:
2580:
2568:
2536:
2523:
2514:
2450:
2434:
2427:
2412:
2397:
2368:
2349:
2334:
2314:
2294:
2270:
2259:
2243:vector space
2236:
2230:
2227:
2198:
2177:
2173:
2165:
2106:Amazon Alexa
2091:
2045:
1915:
1911:
1897:Applications
1880:multiplexing
1866:
1851:
1820:
1807:
1795:
1780:
1713:weight decay
1706:
1699:
1684:
1674:
1670:
1666:
1655:
1651:
1647:
1638:
1619:
1611:
1607:
1603:real numbers
1584:
1563:
1559:
1555:
1554:
1486:Turing Award
1472:
1449:
1438:
1400:
1390:(2015), and
1385:
1374:
1367:
1328:
1313:
1296:
1294:
1283:
1279:
1276:
1251:
1247:
1243:MNIST images
1224:Geoff Hinton
1221:
1202:
1191:
1178:Mel-Cepstral
1171:
1140:
1125:
1090:
1063:
1045:
1032:
1028:
1020:
1002:
984:
949:
910:
902:Neocognitron
895:
880:
857:
845:
835:
813:created the
807:Wilhelm Lenz
788:
736:concepts of
734:optimization
723:
708:
680:
669:
656:Rina Dechter
651:
649:
641:
634:
621:
619:
605:
603:
577:
557:transformers
546:
530:
486:transformers
463:
459:unsupervised
425:
424:
298:Chinese room
187:Applications
153:
15014:Categories
14962:Autoencoder
14917:Transformer
14785:Alex Graves
14733:OpenAI Five
14637:IBM Watsonx
14259:Convolution
14237:Overfitting
14001:22 November
13471:Gary Marcus
12391:10.14336/AD
12173:: 489–504.
11925:: 686–707.
11888:EurekAlert!
11749:VentureBeat
11526:(7): 1819.
11463:IEEE Access
11124:(4): e125.
11097:23 November
10076:21 December
9810:: 333–338.
9540:27 December
9378:VentureBeat
9317:(2): 1–12.
9090:: 197–227.
8978:30 November
8652:Interspeech
8010:25 November
7196:Interspeech
7110:27 December
7042:Sainath, T.
6976:(5): 5947.
6659:: 369–376.
5031:ARS Journal
4839:Robbins, H.
3876:27 December
3798:(5): 5947.
3577:3 September
2958:psychedelic
2858:Gary Marcus
2749:transducers
2415:autoencoder
2341:Ebola virus
2326:off-targets
1702:overfitting
1632:filtering,
1519:sea urchins
1432:(2022) and
1363:Inceptionv3
1353:network by
1309:max-pooling
1205:Alex Graves
1174:filter-bank
1097:Peter Dayan
1033:automatizer
956:Alex Waibel
946:1980s-2000s
936:Paul Werbos
918:derived by
862:trained by
831:Alan Turing
815:Ising model
811:Ernst Ising
785:Before 1980
758:regularizer
537:human brain
514:drug design
327:Turing test
303:Friendly AI
74:Major goals
15003:Technology
14856:EleutherAI
14815:Fei-Fei Li
14810:Yann LeCun
14723:Q-learning
14706:Decisional
14632:IBM Watson
14540:Midjourney
14432:TensorFlow
14279:Activation
14232:Regression
14227:Clustering
14089:2021-05-09
13889:2017-10-11
13850:11 October
13811:11 October
13781:11 October
13752:2015-05-10
13638:2015-05-10
13543:2017-06-14
13512:2 November
13481:11 October
13348:30 January
13255:Huang, Aja
12362:2024-05-19
12303:2212.12794
12237:2006.14395
12180:2008.11625
12111:1707.02568
11990:: 112789.
11754:2023-12-19
11729:2018-07-15
11686:: 105048.
11664:2018-01-01
11629:11 October
11443:2019-11-12
11355:1702.05747
11301:2012.11197
11262:2024-05-10
11238:2024-05-10
10951:2017-06-14
10917:1504.01840
10896:2019-09-05
10823:1704.01212
10801:9 November
10771:9 November
10737:1510.02855
10690:2015-03-05
10626:2020-07-16
10616:kaggle.com
10589:1942/18723
10523:(8): 569.
10487:1 December
10451:12 October
10419:1609.08144
10273:(1): 157.
10040:2014-09-03
10003:26 October
9788:2014-01-28
9706:2019-04-01
9617:2018-04-20
9475:2002.00281
9384:2022-08-03
9359:2022-08-03
9324:1704.04760
9181:23 October
9097:1702.07908
9009:2017-06-13
8949:2017-06-13
8681:2017-06-13
8629:1512.00103
8605:1602.02410
8581:2017-06-13
8550:Proc. NIPS
8522:2020-02-25
8430:2021-01-05
8400:2017-06-13
8371:2020-11-16
8245:2024-08-07
8221:2017-06-13
8208:Google.com
8113:2102.04029
8088:2016-04-09
8047:1503.03585
7985:1710.10196
7963:October 3,
7903:1508.06576
7864:1512.03385
7833:1512.03385
7808:1502.01852
7581:2017-06-13
7547:2017-06-13
7405:1703.09039
7382:2021-02-14
7282:2017-06-14
7236:2017-06-14
7206:2017-06-12
7013:Yann LeCun
6839:2011-07-20
6634:2016-04-09
6472:2017-06-12
6395:2019-09-24
6323:2017-06-12
5921:1906.04493
5545:October 7,
4903:1710.05941
4853:(3): 400.
4824:2019-11-05
4766:Automatica
4506:2212.11279
4391:2017-08-06
4331:1505.03654
4238:2014-10-18
4148:(1): 138.
3769:2019-10-06
3677:: 85–117.
3504:2023-12-06
3468:TechCrunch
3449:2017-05-24
3250:References
3163:) and (5)
3089:April 2021
3059:newspapers
2992:Google Now
2966:stop signs
2838:See also:
2551:inpainting
2465:classifier
2310:Toxicology
2193:The Scream
2114:Apple Siri
2110:Google Now
2063:Multi-task
1888:integrated
1876:wavelength
1864:(FGFETs).
1696:Challenges
1658:primitives
1482:Yann LeCun
1458:(ASR) and
1445:smartphone
1404:(GAN) by (
1239:fine-tuned
1163:Larry Heck
1115:, and the
1025:distilling
980:Yann LeCun
960:Yann LeCun
916:chain rule
801:(MLP) and
606:on its own
583:algorithms
526:board game
451:supervised
332:Regulation
286:Philosophy
241:Healthcare
236:Government
138:Approaches
14886:MIT CSAIL
14851:Anthropic
14820:Andrew Ng
14718:AlphaZero
14562:VideoPoet
14525:AlphaFold
14462:MindSpore
14416:SpiNNaker
14411:Memristor
14318:Diffusion
14294:Rectifier
14274:Batchnorm
14254:Attention
14249:Adversary
14098:cite book
13991:1059-1028
13965:209363848
13957:1461-4448
13704:CiteSeerX
13680:1312.6199
13659:1412.1897
13450:29 August
13297:0028-0836
13178:26 August
13122:1411.6422
13064:0962-8436
13009:1546-1726
12958:0959-4388
12907:0896-6273
12840:1553-7358
12775:2397-3374
12714:1662-5188
12665:0899-7667
12612:0027-8424
12520:CiteSeerX
12330:0036-8075
12270:220055785
12254:1525-8955
12205:235340737
12197:2333-9403
12022:212755458
12014:0045-7825
11949:0021-9991
11898:29 August
11865:265503872
11804:1476-4687
11700:204092079
11542:2072-6694
11501:220733699
11493:2169-3536
11348:: 60–88.
11326:229339809
11197:1067-5027
11087:207217210
10867:201716327
9968:1402.3722
9940:(4): 18.
9854:1404.3840
9812:CiteSeerX
9759:1412.5567
9693:217950236
9607:206602362
9500:211010976
8882:CiteSeerX
8877:1212.0901
8717:0899-7667
8559:1409.3215
8329:1476-4687
8279:cite book
8185:1410.4281
8140:231846518
7937:20 August
7788:1409.1556
7762:1411.2539
7740:1411.4952
7717:1411.4555
7696:1409.4842
7664:1409.1556
7643:1112.6209
7488:0899-7667
7471:1003.0358
7074:206485943
7000:1941-6016
6943:0899-7667
6743:1364-6613
6661:CiteSeerX
6374:0096-3518
6281:0218-0014
6238:0022-2836
5985:0364-0213
5946:216056336
5914:: 58–66.
5796:Q98967430
5609:0364-0213
5518:CiteSeerX
5253:1476-4687
5099:122357351
4986:206775608
4943:206775608
4676:1939-1471
3936:235081987
3928:2405-8963
3684:1404.7828
3616:1206.5538
3564:207178999
3542:CiteSeerX
3384:1202.2745
3296:220523562
3288:1610-1987
3165:clickwork
3125:microwork
3119:(e.g. on
3117:clickwork
2973:deception
2896:website.
2854:strong AI
2847:black box
2788:Google's
2745:neocortex
2696:CpG sites
2543:denoising
2441:AlphaFold
2079:of speech
1891:photonics
1878:division
1857:based on
1755:ℓ
1724:ℓ
1574:manually
1422:deepfakes
1320:Jeff Dean
1316:Andrew Ng
1314:In 2012,
1286:Andrew Ng
1182:waveforms
1064:In 1991,
1021:collapsed
1015:to learn
881:In 1969,
797:(FNN) or
650:The term
614:discovers
571:and deep
362:AI winter
263:Military
126:AI safety
15040:Category
14994:Portals
14753:Auto-GPT
14585:Word2vec
14389:Hardware
14306:Datasets
14208:Concepts
14083:Archived
13995:Archived
13911:: 38–39.
13883:Archived
13844:Archived
13805:Archived
13775:Archived
13629:Archived
13594:Archived
13563:Archived
13537:Archived
13506:Archived
13475:Archived
13444:Archived
13373:Archived
13305:26819042
13229:26819021
13172:Archived
13149:26157000
13090:39281431
13082:29292348
13025:16970545
13017:26906502
12974:16560320
12966:15321069
12923:14663106
12915:10069343
12858:22096452
12791:24504018
12783:31024135
12732:27468262
12542:10097006
12497:12662587
12427:12396572
12338:37962497
12262:32746211
12148:30082389
12079:32001523
11965:57379996
11892:Archived
11857:38030771
11822:38030720
11813:10700131
11723:Archived
11655:Archived
11623:Archived
11601:35350962
11560:35406591
11437:Archived
11380:28778026
11318:36155469
11215:27521897
11150:27815231
11091:Archived
11028:Archived
11005:31127715
10942:Archived
10890:Archived
10859:31477924
10795:Archived
10765:Archived
10684:Archived
10649:Archived
10620:Archived
10598:25582842
10547:20246434
10539:23903212
10445:Archived
10396:23 March
10390:Archived
10358:23 March
10352:Archived
10321:Archived
10299:36855134
10248:40745740
10205:Archived
10132:Archived
10101:Archived
10067:Archived
10031:Archived
9994:Archived
9902:(4): 5.
9834:22386783
9782:Archived
9730:Archived
9697:Archived
9649:15641618
9611:Archived
9492:33408373
9442:33149289
9354:Datanami
9266:Archived
9236:Archived
9205:Archived
9175:Archived
9149:Archived
9122:14135321
9061:Archived
9003:Archived
8972:Archived
8968:Coursera
8940:Archived
8914:12485056
8853:13816461
8805:14542261
8758:Archived
8672:Archived
8668:17048224
8572:Archived
8516:Archived
8512:10192330
8504:18249962
8424:Archived
8394:Archived
8365:archived
8337:26819042
8212:Archived
8082:Archived
8004:Archived
7928:Archived
7687:Cvpr2015
7623:24579167
7572:Archived
7538:Archived
7496:20858131
7376:archived
7306:Archived
7276:Archived
7230:Archived
7200:Archived
7172:16 March
7166:Archived
7101:Archived
7020:Archived
6951:16764513
6904:17921042
6858:Archived
6830:Archived
6818:16764513
6763:Archived
6759:15066318
6751:17921042
6684:Archived
6625:Archived
6588:Archived
6466:Archived
6386:Archived
6317:Archived
6297:(1992).
5938:32334341
5792:Wikidata
5728:Archived
5684:18271205
5540:14542261
5441:20706526
5394:20577468
5308:et al.,
5270:Archived
5170:Archived
5157:(1982).
5059:(1970).
4815:Archived
4703:(1962).
4684:13602029
4382:Archived
4348:12149203
4229:Archived
4182:28743932
4120:Archived
3834:Archived
3760:Archived
3709:11715509
3701:25462637
3633:23787338
3472:Archived
3440:Archived
3349:26017442
3178:See also
3149:Facebook
3137:CAPTCHAs
2907:Goertzel
2779:Facebook
2607:Military
2541:such as
2493:alphabet
2350:In 2017
2282:uses an
2238:word2vec
1918:dialects
1869:photonic
1821:Special
1804:Hardware
1791:batching
1744:sparsity
1523:features
1515:starfish
1436:(2022).
1430:DALL·E 2
1418:StyleGAN
1377:training
1326:videos.
1088:(GANs).
1055:residual
770:Narendra
762:Hopfield
738:training
551:such as
543:Overview
385:Glossary
379:Glossary
357:Progress
352:Timeline
312:Takeover
273:Projects
246:Industry
209:Finance
199:Deepfake
149:Symbolic
121:Robotics
96:Planning
14876:Meta AI
14713:AlphaGo
14697:PanGu-Σ
14667:ChatGPT
14642:Granite
14590:Seq2seq
14569:Whisper
14490:WaveNet
14485:AlexNet
14457:Flux.jl
14437:PyTorch
14289:Sigmoid
14284:Softmax
14149:General
13771:Gizmodo
13600:20 June
13569:20 June
13422:5613334
13277:Bibcode
13237:4460235
13207:Bibcode
13140:6605414
13073:5784047
12866:7504633
12849:3207943
12818:Bibcode
12740:9868901
12723:4943066
12673:2376781
12630:1903542
12590:Bibcode
12550:5818342
12435:1119517
12308:Bibcode
12290:Science
12139:6112690
12116:Bibcode
12070:7219083
12049:Bibcode
12041:Science
11992:Bibcode
11957:1595805
11927:Bibcode
11782:Bibcode
11551:8997449
11520:Cancers
11471:Bibcode
11433:4728736
11388:2088679
11360:Bibcode
11206:5391725
11158:3821594
11141:5116102
11034:14 June
10996:6555124
10715:5 March
10655:14 June
10327:14 June
10290:9972634
10211:14 June
10179:1317136
10138:14 June
10107:14 June
9736:14 June
9433:7116757
9412:Bibcode
9272:11 June
9242:11 June
9211:11 June
9102:Bibcode
9067:5 March
9057:8869270
8764:13 June
8733:1915014
8725:9377276
8564:Bibcode
8460:10 July
8309:Bibcode
7504:1918673
7340:Bibcode
7312:14 June
7054:Bibcode
6978:Bibcode
6826:2309950
6769:12 June
6594:14 June
6419:Bibcode
6382:9563026
6246:3172241
6166:7761831
6146:Bibcode
6138:Science
6105:1890561
6097:7584891
6027:194–281
5738:, 1991.
5488:8058017
5468:Bibcode
5421:Bibcode
5374:Bibcode
5233:Bibcode
4978:7370364
4935:7370364
4566:6953413
4534:Bibcode
4436:Bibcode
4173:5527101
4150:Bibcode
4040:7343126
3993:3958369
3973:Bibcode
3800:Bibcode
3478:17 June
3411:2161592
3357:3074096
3329:Bibcode
3073:scholar
2981:malware
2977:malware
2860:noted:
2798:AlphaGo
2712:obesity
2489:Softmax
2122:iFlyTek
2094:Cortana
1825:called
1775:dropout
1599:synapse
1591:neurons
1576:labeled
1567:systems
1331:AlexNet
1324:YouTube
1041:ChatGPT
1029:chunker
780:History
754:dropout
742:testing
695:sigmoid
660:Boolean
367:AI boom
345:History
268:Physics
14891:Huawei
14871:OpenAI
14773:People
14743:MuZero
14605:Gemini
14600:Claude
14535:DALL-E
14447:Theano
14075:
14047:
14028:
13989:
13963:
13955:
13871:Nature
13706:
13420:
13410:
13379:5 July
13313:515925
13311:
13303:
13295:
13268:Nature
13235:
13227:
13199:Nature
13147:
13137:
13088:
13080:
13070:
13062:
13023:
13015:
13007:
12972:
12964:
12956:
12921:
12913:
12905:
12885:Neuron
12864:
12856:
12846:
12838:
12789:
12781:
12773:
12738:
12730:
12720:
12712:
12698:: 73.
12671:
12663:
12628:
12618:
12610:
12548:
12540:
12522:
12495:
12458:
12433:
12425:
12357:Medium
12336:
12328:
12268:
12260:
12252:
12203:
12195:
12146:
12136:
12077:
12067:
12020:
12012:
11963:
11955:
11947:
11863:
11855:
11841:Nature
11820:
11810:
11802:
11774:Nature
11698:
11599:
11589:
11558:
11548:
11540:
11499:
11491:
11431:
11421:
11386:
11378:
11324:
11316:
11213:
11203:
11195:
11156:
11148:
11138:
11085:
11075:
11003:
10993:
10865:
10857:
10596:
10545:
10537:
10297:
10287:
10246:
10177:
9832:
9814:
9691:
9647:
9605:
9531:
9498:
9490:
9462:Nature
9440:
9430:
9404:Nature
9262:OpenAI
9120:
9055:
9045:
8936:ICASSP
8912:
8902:
8884:
8851:
8841:
8803:
8731:
8723:
8715:
8666:
8510:
8502:
8345:515925
8343:
8335:
8327:
8301:Nature
8267:
8138:
8128:
7879:
7621:
7611:
7502:
7494:
7486:
7439:
7272:398770
7270:
7140:
7072:
7017:Online
6998:
6949:
6941:
6902:
6824:
6816:
6757:
6749:
6741:
6663:
6447:357467
6445:
6380:
6372:
6309:
6303:ICASSP
6279:
6244:
6236:
6195:
6174:871473
6172:
6164:
6103:
6095:
6033:
5983:
5944:
5936:
5887:234198
5885:
5824:
5794:
5760:
5682:
5607:
5538:
5520:
5486:
5439:
5392:
5344:et al.
5342:LeCun
5251:
5225:Nature
5202:
5179:2 July
5139:14 Sep
5097:
5009:
4984:
4976:
4941:
4933:
4743:
4682:
4674:
4595:
4564:
4557:346238
4554:
4374:
4346:
4269:
4180:
4170:
4096:
4066:
4038:
3991:
3934:
3926:
3867:
3831:Online
3732:
3707:
3699:
3641:393948
3639:
3631:
3562:
3544:
3409:
3399:
3355:
3347:
3321:Nature
3294:
3286:
3075:
3068:
3061:
3054:
3046:
2948:TinEye
2929:(AI).
2900:Errors
2874:Watson
2834:Theory
2553:, and
2485:pixels
2473:matrix
2469:vector
2432:data.
2328:, and
2126:Nuance
1939:Method
1930:bigram
1831:Huawei
1815:OpenAI
1414:Nvidia
1396:VGG-19
1351:VGG-16
1341:, and
1303:, and
1297:DanNet
1037:layers
774:Bishop
766:Widrow
637:greedy
599:pixels
595:tensor
488:, and
317:Ethics
14957:Mamba
14728:SARSA
14692:LLaMA
14687:BLOOM
14672:GPT-J
14662:GPT-4
14657:GPT-3
14652:GPT-2
14647:GPT-1
14610:LaMDA
14442:Keras
13983:Wired
13961:S2CID
13675:arXiv
13654:arXiv
13632:(PDF)
13625:(PDF)
13418:S2CID
13309:S2CID
13233:S2CID
13168:Wired
13117:arXiv
13086:S2CID
13021:S2CID
12970:S2CID
12919:S2CID
12862:S2CID
12787:S2CID
12736:S2CID
12669:S2CID
12621:51674
12546:S2CID
12431:S2CID
12298:arXiv
12266:S2CID
12232:arXiv
12201:S2CID
12175:arXiv
12106:arXiv
12018:S2CID
11961:S2CID
11861:S2CID
11696:S2CID
11658:(PDF)
11651:(PDF)
11597:S2CID
11497:S2CID
11429:S2CID
11384:S2CID
11350:arXiv
11322:S2CID
11296:arXiv
11284:(PDF)
11154:S2CID
11083:S2CID
10945:(PDF)
10938:(PDF)
10912:arXiv
10886:Wired
10863:S2CID
10818:arXiv
10732:arXiv
10543:S2CID
10481:(PDF)
10474:(PDF)
10440:Wired
10414:arXiv
10244:S2CID
10175:S2CID
10070:(PDF)
10059:(PDF)
10034:(PDF)
10023:(PDF)
9997:(PDF)
9990:(PDF)
9963:arXiv
9849:arXiv
9754:arXiv
9726:Wired
9700:(PDF)
9689:S2CID
9667:(PDF)
9645:S2CID
9603:S2CID
9496:S2CID
9470:arXiv
9319:arXiv
9232:ZDNet
9118:S2CID
9092:arXiv
9053:S2CID
8943:(PDF)
8932:(PDF)
8910:S2CID
8872:arXiv
8849:S2CID
8801:S2CID
8729:S2CID
8675:(PDF)
8664:S2CID
8648:(PDF)
8624:arXiv
8600:arXiv
8575:(PDF)
8554:arXiv
8546:(PDF)
8508:S2CID
8341:S2CID
8215:(PDF)
8204:(PDF)
8180:arXiv
8165:(PDF)
8158:(PDF)
8136:S2CID
8108:arXiv
8042:arXiv
8030:(PDF)
7980:arXiv
7931:(PDF)
7924:(PDF)
7898:arXiv
7859:arXiv
7828:arXiv
7803:arXiv
7783:arXiv
7757:arXiv
7735:arXiv
7712:arXiv
7691:arXiv
7683:(PDF)
7659:arXiv
7638:arXiv
7575:(PDF)
7568:(PDF)
7541:(PDF)
7522:(PDF)
7500:S2CID
7466:arXiv
7400:arXiv
7268:S2CID
7104:(PDF)
7097:(PDF)
7070:S2CID
6833:(PDF)
6822:S2CID
6792:(PDF)
6755:S2CID
6628:(PDF)
6617:(PDF)
6443:S2CID
6389:(PDF)
6378:S2CID
6342:(PDF)
6170:S2CID
6101:S2CID
6017:(PDF)
5942:S2CID
5916:arXiv
5883:S2CID
5704:(PDF)
5680:S2CID
5660:(PDF)
5635:(PDF)
5536:S2CID
5506:(PDF)
5293:(PDF)
5173:(PDF)
5162:(PDF)
5095:S2CID
4982:S2CID
4939:S2CID
4898:arXiv
4818:(PDF)
4799:(PDF)
4501:arXiv
4385:(PDF)
4368:(PDF)
4344:S2CID
4326:arXiv
4232:(PDF)
4209:(PDF)
4036:S2CID
4000:(PDF)
3989:S2CID
3959:(PDF)
3932:S2CID
3763:(PDF)
3756:(PDF)
3705:S2CID
3679:arXiv
3637:S2CID
3611:arXiv
3571:(PDF)
3560:S2CID
3530:(PDF)
3443:(PDF)
3432:(PDF)
3407:S2CID
3379:arXiv
3353:S2CID
3317:(PDF)
3292:S2CID
3080:JSTOR
3066:books
2794:Atari
2118:Baidu
2040:16.5
2032:17.8
2024:18.3
2016:18.7
2008:20.0
2000:20.7
1992:21.7
1984:22.4
1976:23.4
1968:24.8
1960:25.6
1952:26.1
1926:phone
1593:in a
1562:) or
1540:shell
1460:MNIST
1456:TIMIT
1255:TIMIT
1188:2000s
1155:DARPA
1141:Most
976:LeNet
964:LeNet
791:types
591:image
436:with
229:Music
224:Audio
14881:Mila
14682:PaLM
14615:Bard
14595:BERT
14578:Text
14557:Sora
14104:link
14073:ISBN
14045:ISBN
14026:ISBN
14003:2019
13987:ISSN
13953:ISSN
13852:2017
13813:2019
13783:2019
13602:2015
13571:2015
13514:2017
13483:2018
13452:2018
13408:ISBN
13381:2019
13350:2016
13301:PMID
13293:ISSN
13225:PMID
13180:2017
13145:PMID
13078:PMID
13060:ISSN
13013:PMID
13005:ISSN
12962:PMID
12954:ISSN
12911:PMID
12903:ISSN
12854:PMID
12836:ISSN
12779:PMID
12771:ISSN
12728:PMID
12710:ISSN
12661:ISSN
12626:PMID
12608:ISSN
12538:PMID
12493:PMID
12456:ISBN
12423:PMID
12334:PMID
12326:ISSN
12258:PMID
12250:ISSN
12193:ISSN
12144:PMID
12075:PMID
12010:ISSN
11953:OSTI
11945:ISSN
11900:2018
11853:PMID
11818:PMID
11800:ISSN
11631:2019
11587:ISBN
11556:PMID
11538:ISSN
11489:ISSN
11419:ISBN
11376:PMID
11314:PMID
11257:CNBC
11211:PMID
11193:ISSN
11146:PMID
11099:2015
11073:ISBN
11036:2017
11001:PMID
10855:PMID
10803:2015
10773:2015
10717:2015
10657:2017
10594:PMID
10535:PMID
10489:2016
10453:2017
10398:2017
10360:2017
10329:2017
10295:PMID
10213:2017
10140:2017
10109:2017
10078:2023
10005:2014
9934:Arts
9896:Arts
9830:PMID
9738:2017
9675:2015
9542:2023
9529:ISBN
9488:PMID
9438:PMID
9274:2020
9244:2020
9213:2020
9183:2015
9069:2018
9043:ISBN
8980:2017
8900:ISBN
8839:ISBN
8766:2017
8721:PMID
8713:ISSN
8500:PMID
8462:2018
8333:PMID
8325:ISSN
8285:link
8265:ISBN
8126:ISBN
8012:2020
7965:2019
7939:2019
7877:ISBN
7619:PMID
7609:ISBN
7492:PMID
7484:ISSN
7437:ISBN
7314:2017
7174:2018
7138:ISBN
7112:2023
6996:ISSN
6947:PMID
6939:ISSN
6900:PMID
6814:PMID
6771:2017
6747:PMID
6739:ISSN
6596:2017
6370:ISSN
6307:ISBN
6277:ISSN
6242:PMID
6234:ISSN
6193:ISBN
6162:PMID
6093:PMID
6031:ISBN
5999:link
5981:ISSN
5934:PMID
5822:ISBN
5758:ISBN
5605:ISSN
5547:2016
5484:PMID
5437:PMID
5390:PMID
5249:ISSN
5200:ISBN
5181:2017
5141:2024
5007:ISBN
4974:PMID
4931:PMID
4741:ISBN
4680:PMID
4672:ISSN
4593:ISBN
4562:PMID
4372:ISBN
4267:ISBN
4178:PMID
4094:ISBN
4064:ISBN
3924:ISSN
3878:2023
3865:ISBN
3730:ISBN
3697:PMID
3629:PMID
3579:2015
3480:2018
3397:ISBN
3345:PMID
3284:ISSN
3052:news
2997:In "
2762:and
2585:and
2343:and
2308:and
2120:and
2098:Xbox
2073:CNNs
2065:and
1560:ANNs
1517:and
1480:and
1468:LSTM
1357:and
1318:and
1196:and
1167:NIST
1153:and
966:for
950:The
887:ReLU
809:and
768:and
740:and
724:The
715:ReLU
693:for
555:and
14622:NMT
14505:OCR
14500:HWR
14452:JAX
14406:VPU
14401:TPU
14396:IPU
14220:SGD
13943:doi
13875:doi
13714:doi
13400:doi
13285:doi
13273:529
13215:doi
13203:529
13135:PMC
13127:doi
13068:PMC
13052:doi
13048:373
12997:doi
12946:doi
12893:doi
12844:PMC
12826:doi
12763:doi
12718:PMC
12700:doi
12653:doi
12616:PMC
12598:doi
12530:doi
12485:doi
12415:doi
12386:doi
12316:doi
12294:382
12242:doi
12185:doi
12134:PMC
12124:doi
12102:115
12065:PMC
12057:doi
12045:367
12000:doi
11988:360
11935:doi
11923:378
11845:doi
11808:PMC
11790:doi
11778:624
11688:doi
11684:188
11579:doi
11546:PMC
11528:doi
11479:doi
11411:doi
11368:doi
11306:doi
11201:PMC
11185:doi
11136:PMC
11126:doi
11065:hdl
11057:doi
10991:PMC
10981:doi
10847:doi
10584:hdl
10574:doi
10525:doi
10285:PMC
10275:doi
10236:doi
10167:doi
9942:doi
9904:doi
9822:doi
9679:doi
9637:doi
9595:doi
9564:doi
9521:doi
9480:doi
9466:589
9428:PMC
9420:doi
9408:587
9329:doi
9147:".
9110:doi
9035:doi
8892:doi
8831:doi
8793:doi
8705:doi
8656:doi
8492:doi
8317:doi
8305:529
8118:doi
7869:doi
7601:doi
7530:doi
7476:doi
7429:doi
7348:doi
7260:doi
7062:doi
6986:doi
6931:doi
6892:doi
6856:".
6806:doi
6731:doi
6553:doi
6523:doi
6496:doi
6435:hdl
6427:doi
6362:hdl
6354:doi
6269:doi
6226:doi
6222:202
6154:doi
6142:268
6083:hdl
6075:doi
5973:doi
5926:doi
5912:127
5875:doi
5814:doi
5726:".
5672:doi
5597:doi
5528:doi
5476:doi
5429:doi
5382:doi
5241:doi
5229:323
5087:doi
5039:doi
4966:doi
4923:doi
4855:doi
4807:doi
4774:doi
4664:doi
4618:doi
4585:doi
4552:PMC
4542:doi
4444:doi
4336:doi
4299:doi
4221:doi
4168:PMC
4158:doi
4028:doi
3981:doi
3914:doi
3857:doi
3808:doi
3689:doi
3621:doi
3552:doi
3389:doi
3337:doi
3325:521
3276:doi
3147:on
3035:by
2942:".
2880:to
2700:IBD
2629:CFD
2471:or
2413:An
2378:RFM
2264:to
2159:in
2083:RNN
1920:of
1443:on
1424:.
1416:'s
1333:by
1232:Teh
1151:NSA
756:as
702:'s
674:or
597:of
457:or
219:Art
15042::
14100:}}
14096:{{
14081:.
14061:;
13993:.
13985:.
13981:.
13959:.
13951:.
13939:22
13937:.
13933:.
13917:^
13907:.
13881:.
13873:.
13869:.
13838:.
13821:^
13799:.
13769:.
13712:.
13698:.
13627:.
13610:^
13592:.
13588:.
13535:.
13531:.
13504:.
13500:.
13473:.
13469:.
13438:.
13416:.
13406:.
13371:.
13367:.
13336:.
13307:.
13299:.
13291:.
13283:.
13271:.
13253:;
13231:.
13223:.
13213:.
13201:.
13197:.
13170:.
13166:.
13143:.
13133:.
13125:.
13113:35
13111:.
13107:.
13084:.
13076:.
13066:.
13058:.
13046:.
13042:.
13019:.
13011:.
13003:.
12993:19
12991:.
12968:.
12960:.
12952:.
12942:14
12940:.
12917:.
12909:.
12901:.
12889:22
12887:.
12883:.
12860:.
12852:.
12842:.
12834:.
12824:.
12812:.
12808:.
12785:.
12777:.
12769:.
12757:.
12734:.
12726:.
12716:.
12708:.
12696:10
12694:.
12690:.
12667:.
12659:.
12647:.
12624:.
12614:.
12606:.
12596:.
12586:88
12584:.
12580:.
12544:.
12536:.
12528:.
12516:20
12514:.
12491:.
12479:.
12429:.
12421:.
12411:14
12409:.
12384:.
12380:.
12355:.
12332:.
12324:.
12314:.
12306:.
12292:.
12288:.
12264:.
12256:.
12248:.
12240:.
12228:67
12226:.
12222:.
12199:.
12191:.
12183:.
12169:.
12165:.
12142:.
12132:.
12122:.
12114:.
12100:.
12096:.
12073:.
12063:.
12055:.
12043:.
12039:.
12016:.
12008:.
11998:.
11986:.
11982:.
11959:.
11951:.
11943:.
11933:.
11921:.
11917:.
11890:.
11886:.
11873:^
11859:.
11851:.
11843:.
11839:.
11816:.
11806:.
11798:.
11788:.
11776:.
11772:.
11747:.
11721:.
11717:.
11694:.
11682:.
11617:.
11595:.
11585:.
11554:.
11544:.
11536:.
11524:14
11522:.
11518:.
11495:.
11487:.
11477:.
11465:.
11461:.
11435:.
11427:.
11417:.
11405:.
11382:.
11374:.
11366:.
11358:.
11346:42
11344:.
11320:.
11312:.
11304:.
11292:PP
11290:.
11286:.
11271:^
11231:.
11209:.
11199:.
11191:.
11181:24
11179:.
11175:.
11152:.
11144:.
11134:.
11120:.
11116:.
11089:.
11081:.
11071:.
11063:.
11026:.
11022:.
10999:.
10989:.
10977:21
10975:.
10969:.
10888:.
10884:.
10861:.
10853:.
10843:37
10841:.
10789:.
10763:.
10757:.
10746:^
10682:.
10647:.
10643:.
10618:.
10614:.
10592:.
10582:.
10570:20
10568:.
10564:.
10541:.
10533:.
10521:12
10519:.
10515:.
10461:^
10443:.
10437:.
10388:.
10384:.
10368:^
10350:.
10346:.
10319:.
10315:.
10293:.
10283:.
10271:21
10269:.
10265:.
10242:.
10232:30
10230:.
10203:.
10199:.
10187:^
10173:.
10163:23
10161:.
10130:.
10126:.
10099:.
10095:.
10061:.
10029:.
10025:.
9992:.
9977:^
9936:.
9932:.
9918:^
9898:.
9894:.
9880:^
9828:.
9820:.
9808:32
9780:.
9776:.
9728:.
9724:.
9695:.
9687:.
9677:.
9673:.
9669:.
9643:.
9609:.
9601:.
9591:22
9589:.
9585:.
9527:.
9494:.
9486:.
9478:.
9464:.
9450:^
9436:.
9426:.
9418:.
9406:.
9402:.
9376:.
9352:.
9327:.
9315:45
9313:.
9309:.
9290:.
9260:.
9234:.
9230:.
9199:.
9173:.
9169:.
9116:.
9108:.
9100:.
9088:75
9086:.
9059:.
9051:.
9041:.
9027:.
9001:.
8997:.
8970:.
8966:.
8938:.
8934:.
8908:.
8898:.
8890:.
8880:.
8847:.
8837:.
8799:.
8789:86
8787:.
8783:.
8756:.
8752:.
8741:^
8727:.
8719:.
8711:.
8699:.
8670:.
8662:.
8650:.
8614:^
8590:^
8570:.
8562:.
8552:.
8548:.
8531:^
8514:.
8506:.
8498:.
8488:12
8486:.
8482:.
8470:^
8448:.
8418:.
8388:.
8363:,
8339:.
8331:.
8323:.
8315:.
8303:.
8281:}}
8277:{{
8238:.
8206:.
8134:.
8124:.
8116:.
8080:.
8068:^
8038:37
8036:.
8032:.
7955:.
7875:.
7867:.
7853:.
7826:.
7781:,
7689:.
7685:.
7617:.
7607:.
7536:.
7528:.
7524:.
7498:.
7490:.
7482:.
7474:.
7462:22
7460:.
7435:.
7423:.
7374:,
7360:^
7346:.
7336:37
7334:.
7322:^
7304:.
7300:.
7274:.
7266:.
7254:.
7228:.
7224:.
7198:.
7194:.
7182:^
7160:.
7120:^
7082:^
7068:.
7060:.
7050:29
7048:.
7030:^
6994:.
6984:.
6972:.
6968:.
6945:.
6937:.
6927:18
6925:.
6921:.
6898:.
6888:11
6886:.
6882:.
6864:.
6828:.
6820:.
6812:.
6802:18
6800:.
6794:.
6761:.
6753:.
6745:.
6737:.
6727:11
6725:.
6721:.
6619:.
6604:^
6586:.
6582:.
6549:31
6547:.
6535:^
6519:31
6517:.
6490:.
6441:.
6433:.
6425:.
6415:26
6413:.
6384:.
6376:.
6368:.
6360:.
6350:37
6348:.
6344:.
6315:.
6301:.
6275:.
6265:07
6263:.
6240:.
6232:.
6220:.
6216:.
6168:.
6160:.
6152:.
6140:.
6132:;
6128:;
6099:.
6091:.
6081:.
6069:.
6061:;
6057:;
6053:;
6029:.
5995:}}
5991:{{
5967:.
5963:.
5940:.
5932:.
5924:.
5910:.
5895:^
5881:.
5869:.
5836:^
5820:.
5790:,
5780:;
5734:.
5713:^
5678:.
5666:.
5662:.
5647:^
5637:.
5617:^
5603:.
5593:14
5591:.
5587:.
5568:.
5564:.
5534:.
5526:.
5514:86
5512:.
5508:.
5482:.
5474:.
5464:21
5462:.
5458:.
5435:.
5427:.
5417:30
5415:.
5411:.
5388:.
5380:.
5370:29
5368:.
5364:.
5328:.
5247:.
5239:.
5227:.
5223:.
5164:.
5116:^
5093:.
5083:16
5081:.
5035:30
5033:.
4980:.
4972:.
4962:36
4960:.
4937:.
4929:.
4883:EC
4881:.
4851:22
4849:.
4845:.
4813:.
4801:.
4786:^
4768:.
4764:.
4692:^
4678:.
4670:.
4660:65
4658:.
4654:.
4591:.
4560:.
4550:.
4540:.
4530:79
4528:.
4524:.
4478:^
4468:.
4456:^
4442:.
4432:39
4430:.
4411:.
4400:^
4380:.
4342:.
4334:.
4322:43
4320:.
4293:.
4281:^
4247:^
4227:.
4215:.
4211:.
4190:^
4176:.
4166:.
4156:.
4144:.
4140:.
4108:^
4078:^
4048:^
4034:.
4022:.
4008:^
3987:.
3979:.
3967:.
3961:.
3944:^
3930:.
3922:.
3910:53
3904:.
3863:.
3806:.
3794:.
3790:.
3778:^
3703:.
3695:.
3687:.
3675:61
3673:.
3649:^
3635:.
3627:.
3619:.
3607:35
3605:.
3587:^
3558:.
3550:.
3536:.
3532:.
3513:^
3496:.
3466:.
3438:.
3434:.
3419:^
3405:.
3395:.
3387:.
3365:^
3351:.
3343:.
3335:.
3323:.
3319:.
3304:^
3290:.
3282:.
3272:26
3270:.
3266:.
3167:.
2894:'s
2884:.
2802:Go
2710:,
2706:,
2702:,
2549:,
2545:,
2384:.
2347:.
2324:,
2268:.
2116:,
2112:,
2108:,
2104:,
2100:,
2096:,
1845:.
1628:,
1624:,
1582:.
1476:,
1447:.
1398:.
1365:.
1337:,
1261:.
1226:,
1219:.
1111:,
1107:,
1099:,
1095:,
1000:.
776:.
764:,
706:.
678:.
647:.
575:.
520:,
516:,
512:,
508:,
504:,
500:,
496:,
484:,
480:,
476:,
472:,
468:,
461:.
453:,
14135:e
14128:t
14121:v
14106:)
14053:.
14034:.
14005:.
13967:.
13945::
13892:.
13877::
13854:.
13815:.
13785:.
13755:.
13720:.
13716::
13700:2
13683:.
13677::
13662:.
13656::
13641:.
13604:.
13573:.
13546:.
13516:.
13485:.
13454:.
13424:.
13402::
13383:.
13352:.
13315:.
13287::
13279::
13239:.
13217::
13209::
13182:.
13151:.
13129::
13119::
13092:.
13054::
13027:.
12999::
12976:.
12948::
12925:.
12895::
12868:.
12828::
12820::
12814:7
12793:.
12765::
12759:1
12742:.
12702::
12675:.
12655::
12649:8
12632:.
12600::
12592::
12552:.
12532::
12499:.
12487::
12481:9
12464:.
12437:.
12417::
12394:.
12388::
12365:.
12340:.
12318::
12310::
12300::
12272:.
12244::
12234::
12207:.
12187::
12177::
12171:7
12150:.
12126::
12118::
12108::
12081:.
12059::
12051::
12024:.
12002::
11994::
11967:.
11937::
11929::
11902:.
11867:.
11847::
11824:.
11792::
11784::
11757:.
11732:.
11702:.
11690::
11667:.
11633:.
11603:.
11581::
11562:.
11530::
11503:.
11481::
11473::
11467:8
11446:.
11413::
11390:.
11370::
11362::
11352::
11328:.
11308::
11298::
11265:.
11241:.
11217:.
11187::
11160:.
11128::
11122:4
11101:.
11067::
11059::
11038:.
11007:.
10983::
10954:.
10920:.
10914::
10899:.
10869:.
10849::
10826:.
10820::
10805:.
10775:.
10740:.
10734::
10719:.
10693:.
10659:.
10629:.
10600:.
10586::
10576::
10549:.
10527::
10491:.
10455:.
10422:.
10416::
10400:.
10362:.
10331:.
10301:.
10277::
10250:.
10238::
10215:.
10181:.
10169::
10142:.
10111:.
10080:.
10043:.
10007:.
9971:.
9965::
9950:.
9944::
9938:6
9912:.
9906::
9900:6
9857:.
9851::
9836:.
9824::
9791:.
9762:.
9756::
9740:.
9709:.
9681::
9651:.
9639::
9620:.
9597::
9570:.
9566::
9544:.
9523::
9502:.
9482::
9472::
9444:.
9422::
9414::
9387:.
9362:.
9337:.
9331::
9321::
9294:.
9276:.
9246:.
9215:.
9185:.
9124:.
9112::
9104::
9094::
9071:.
9037::
9012:.
8982:.
8952:.
8916:.
8894::
8874::
8855:.
8833::
8807:.
8795::
8768:.
8735:.
8707::
8701:9
8684:.
8658::
8632:.
8626::
8608:.
8602::
8584:.
8566::
8556::
8525:.
8494::
8464:.
8433:.
8403:.
8347:.
8319::
8311::
8287:)
8273:.
8248:.
8224:.
8188:.
8182::
8142:.
8120::
8110::
8091:.
8050:.
8044::
8014:.
7988:.
7982::
7967:.
7941:.
7906:.
7900::
7885:.
7871::
7861::
7836:.
7830::
7811:.
7805::
7785::
7767:.
7765:.
7759::
7745:.
7743:.
7737::
7722:.
7720:.
7714::
7699:.
7693::
7667:.
7661::
7646:.
7640::
7625:.
7603::
7584:.
7550:.
7532::
7506:.
7478::
7468::
7445:.
7431::
7408:.
7402::
7354:.
7350::
7342::
7316:.
7285:.
7262::
7239:.
7209:.
7176:.
7146:.
7114:.
7076:.
7064::
7056::
7002:.
6988::
6980::
6974:4
6953:.
6933::
6906:.
6894::
6842:.
6808::
6773:.
6733::
6669:.
6637:.
6598:.
6559:.
6555::
6529:.
6525::
6502:.
6498::
6492:7
6475:.
6449:.
6437::
6429::
6421::
6398:.
6364::
6356::
6326:.
6283:.
6271::
6248:.
6228::
6201:.
6176:.
6156::
6148::
6107:.
6085::
6077::
6071:7
6039:.
6001:)
5987:.
5975::
5969:9
5948:.
5928::
5918::
5889:.
5877::
5871:2
5830:.
5816::
5766:.
5706:.
5686:.
5674::
5668:4
5641:.
5611:.
5599::
5572:.
5570:8
5549:.
5530::
5490:.
5478::
5470::
5443:.
5431::
5423::
5396:.
5384::
5376::
5332:.
5255:.
5243::
5235::
5208:.
5183:.
5143:.
5101:.
5089::
5045:.
5041::
5015:.
4988:.
4968::
4945:.
4925::
4906:.
4900::
4863:.
4857::
4827:.
4809::
4780:.
4776::
4770:6
4749:.
4722:.
4686:.
4666::
4639:.
4624:.
4620::
4601:.
4587::
4568:.
4544::
4536::
4509:.
4503::
4470:C
4450:.
4446::
4438::
4415:.
4394:.
4350:.
4338::
4328::
4305:.
4301::
4295:5
4275:.
4241:.
4223::
4217:7
4184:.
4160::
4152::
4146:8
4102:.
4072:.
4042:.
4030::
4024:4
3983::
3975::
3969:2
3938:.
3916::
3880:.
3859::
3816:.
3810::
3802::
3796:4
3772:.
3738:.
3711:.
3691::
3681::
3643:.
3623::
3613::
3581:.
3554::
3538:2
3507:.
3482:.
3452:.
3413:.
3391::
3381::
3359:.
3339::
3331::
3298:.
3278::
3102:)
3096:(
3091:)
3087:(
3077:·
3070:·
3063:·
3056:·
3029:.
2952:.
2312:.
1759:1
1746:(
1728:2
1715:(
1558:(
1462:(
1134:/
414:e
407:t
400:v
310:/
34:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.