Deep learning: Difference between revisions

Revision as of 18:19, 9 April 2016

Revision as of 18:58, 9 April 2016

Line 25:

These definitions have in common (1) multiple layers of nonlinear processing units and (2) the supervised or unsupervised learning of feature representations in each layer, with the layers forming a hierarchy from low-level to high-level features.<ref name="BOOK2014" />{{rp|page=200}} The composition of a layer of nonlinear processing units used in a deep learning algorithm depends on the problem to be solved. Layers that have been used in deep learning include hidden layers of an [[artificial neural network]] and sets of complicated [[propositional formula]]s.<ref name="BENGIODEEP" /> They may also include latent variables organized layer-wise in deep generative models such as the nodes in Deep Belief Networks and Deep Boltzmann Machines.

Deep learning algorithms transform their inputs through more layers than shallow learning algorithms. At each layer, the signal is transformed by a processing unit, like an artificial neuron, whose parameters are ‘learned’'learned' through training.<ref name="SCHIDHUB" />{{rp|page=6}} A chain of transformations from input to output is a ''credit assignment path'' (CAP). CAPs describe potentially causal connections between input and output and may vary in length. For a feedforward neural network, the depth of the CAPs, and thus the depth of the network, is the number of hidden layers plus one (the output layer is also parameterized). For [[recurrent neural network]]s, in which a signal may propagate through a layer more than once, the CAP is potentially unlimited in length. There is no universally agreed upon threshold of depth dividing shallow learning from deep learning, but most researchers in the field agree that deep learning has multiple nonlinear layers (CAP > 2) and Schmidhuber considers CAP > 10 to be very deep learning.<ref name="SCHIDHUB" />{{rp|page=7}}

=== Fundamental concepts ===

Line 84:

== Artificial neural networks ==

Some of the most successful deep learning methods involve artificial [[neural network]]s. Artificial neural networks are inspired by the 1959 biological model proposed by [[Nobel laureate]]s [[David H. Hubel]] & [[Torsten Wiesel]], who found two types of cells in the [[primary visual cortex]]: [[simple cell]]s and [[complex cell]]s. Many artificial neural networks can be viewed as cascading models <ref name="Weng1992"/><ref name="Weng1993"/><ref name="Weng1997"/><ref>{{cite journal | last1 = Riesenhuber | first1 = M | last2 = Poggio | first2 = T | year = 1999 | title = Hierarchical models of object recognition in cortex | doi = 10.1038/14819 | journal = Nature Neuroscience | volume = 2 | issue = 11| pages = 1019–1025 }}</ref> of cell types inspired by these biological observations.

Fukushima's Neocognitron introduced [[convolution]]al neural networks partially trained by [[unsupervised learning]] with human-directed features in the neural plane. [[Yann LeCun]] et al. (1989) applied supervised [[backpropagation]] to such architectures.<ref name=LeCun1989>Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, L. D. Jackel. 1989 ''Backpropagation Applied to Handwritten Zip Code Recognition.'' Neural Computation, 1(4):541–551.</ref> Weng et al. (1992) published convolutional neural networks Cresceptron<ref name="Weng1992"/><ref name="Weng1993"/><ref name="Weng1997"/> for 3-D object recognition from images of cluttered scenes and segmentation of such objects from images.

Line 191:

or [[weight decay]] (<math> \ell_2 </math>-regularization) or [[sparse matrix|sparsity]] (<math> \ell_1 </math>-regularization) can be applied during training to help combat overfitting.<ref name="BENGIO2013">Y. Bengio ''et al.''., "Advances in optimizing recurrent networks," ''ICASSP'', 2013.</ref> A more recent regularization method applied to DNNs is [[Dropout (neural networks)|dropout]] regularization. In dropout, some number of units are randomly omitted from the hidden layers during training. This helps to break the rare dependencies that can occur in the training data.<ref name="DAHL2013">G. Dahl ''et al.''., "Improving DNNs for LVCSR using rectified linear units and dropout," ''ICASSP'', 2013.</ref>

The dominant method for training these structures has been error-correction training (such as [[backpropagation]] with [[gradient descent]]) due to its ease of implementation and its tendency to converge to better [[Local optimum|local optima]] than other training methods. However, these methods can be computationally expensive, especially for DNNs. There are many training parameters to be considered with a DNN, such as the size (number of layers and number of units per layer), the learning rate and initial weights. [[Hyperparameter optimization#Grid search|Sweeping through the parameter space]] for optimal parameters may not be feasible due to the cost in time and computational resources. Various 'tricks' such as using mini-batching (computing the gradient on several training examples at once rather than individual examples) <ref name="RBMTRAIN">G. E. Hinton., "A Practical Guide to Training Restricted Boltzmann Machines," ''Tech. Rep. UTML TR 2010-003, Dept. CS., Univ. of Toronto'', 2010.</ref> have been shown to speed up computation. The large processing throughput of GPUs has produced significant speedups in training, due to the matrix and vector computations required being well suited for GPUs.<ref name="SCHIDHUB" /> Radical alternatives to backprop such as [[Extreme Learning Machines]],<ref>{{cite journal | last1 = Huang | first1 = Guang-Bin | last2 = Zhu | first2 = Qin-Yu | last3 = Siew | first3 = Chee-Kheong | year = 2006 | title = Extreme learning machine: theory and applications | url = | journal = Neurocomputing | volume = 70 | issue = 1| pages = 489–501 | doi=10.1016/j.neucom.2005.12.126}}</ref> "No-prop" networks,<ref>{{cite journal | last1 = Widrow | first1 = Bernard | display-authors = etal | year = 2013 | title = The no-prop algorithm: A new learning algorithm for multilayer neural networks | url = | journal = Neural Networks | volume = 37 | issue = | pages = 182–188 | doi=10.1016/j.neunet.2012.09.020}}</ref> training without backtracking,<ref>{{cite arXiv |last=Ollivier |first=Yann |last2=Charpiat |first2=Guillaume |year=2015 |title=Training recurrent networks without backtracking |arxiv=1507.07680}}</ref> "weightless" networks,<ref>Aleksander, Igor, et al. "A brief introduction to Weightless Neural Systems." ESANN. 2009.</ref> and [[holographic associative memory|non-connectionist neural networks]] are gaining attention.

=== First deep learning networks of 1965: GMDH ===

Line 253:

=== Convolutional deep belief networks ===

A recent achievement in deep learning is the use of convolutional deep belief networks (CDBN). CDBNs have structure very similar to a [[convolutional neural network]]s and are trained similar to deep belief networks. Therefore, they exploit the 2D structure of images, like CNNs do, and make use of pre-training like [[deep belief network]]s. They provide a generic structure which can be used in many image and signal processing tasks. Recently, many benchmark results on standard image datasets like CIFAR <ref name="CDBN-CIFAR">[http://www.cs.toronto.edu/~kriz/conv-cifar10-aug2010.pdf Convolutional Deep Belief Networks on CIFAR-10]</ref> have been obtained using CDBNs.<ref name="CDBN">[http://dl.acm.org/citation.cfm?id=1553453 Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations]</ref>

=== Large memory storage and retrieval neural networks ===

Line 569:

| pages = 1096–1103

| date = 2008

| url = http://portal.acm.org/citation.cfm?doid=1390156.1390294}}</ref> This provides a better representation, allowing faster learning and more accurate classification with high-dimensional data. However, these architectures are poor at learning novel classes with few examples, because all network units are involved in representing the input (a ''distributed representation'') and must be adjusted together (high [[degree of freedom]]). Limiting the degree of freedom reduces the number of parameters to learn, facilitating learning of new classes from few examples. [[Hierarchical Bayesian model|''Hierarchical Bayesian (HB)'' models]] allow learning from few examples, for example <ref name="ref34">{{cite journal

| last1 = Kemp

| first1 = Charles

Line 680:

===Networks with seperate memory structures===

Integrating external memory with [[artificial neural networks]] dates to early research in [[distributed representations]] <ref name="Hinton, Geoffrey E 1984">Hinton, Geoffrey E. "Distributed representations." (1984)</ref> and [[Teuvo Kohonen]]'s [[self-organizing map]]s. For example, in [[sparse distributed memory]] or [[hierarchical temporal memory]], the patterns encoded by neural networks are used as addresses for [[content-addressable memory]], with "neurons" essentially serving as address [[encoder]]s and [[Binary decoder|decoder]]s. However, the early controllers of such memories were not differentiable.

====LSTM-related differentiable memory structures====

Line 686:

Apart form [[long short-term memory]] (LSTM), other approaches of the 1990s and 2000s also added differentiable memory to recurrent functions. For example:

*Differentiable push and pop actions for alternative memory networks called ''neural stack machines''<ref name="S. Das, C.L. Giles p. 79">S. Das, C.L. Giles, G.Z. Sun, "Learning Context Free Grammars: Limitations of a Recurrent Neural Network with an External Stack Memory," Proc. 14th Annual Conf. of the Cog. Sci. Soc., p. 79, 1992.</ref><ref name="Mozer, M. C. 1993 pp. 863-870">Mozer, M. C., & Das, S. (1993). A connectionist symbol manipulator that discovers the structure of context-free languages. NIPS 5 (pp. 863-870).</ref>

*Memory networks where the control network's external differentiable storage is in the fast weights of another network <ref name="ReferenceC">{{cite journal | last1 = Schmidhuber | first1 = J. | year = 1992 | title = Learning to control fast-weight memories: An alternative to recurrent nets | url = | journal = Neural Computation | volume = 4 | issue = 1| pages = 131–139 | doi=10.1162/neco.1992.4.1.131}}</ref>

*LSTM ''"forget gates"'' <ref name="F. Gers, N. Schraudolph 2002">{{cite journal | last1 = Gers | first1 = F. | last2 = Schraudolph | first2 = N. | last3 = Schmidhuber | first3 = J. | year = 2002 | title = Learning precise timing with LSTM recurrent networks | url = | journal = JMLR | volume = 3 | issue = | pages = 115–143 }}</ref>

*Self-referential [[recurrent neural networks]] (RNNs) with special output units for addressing and rapidly manipulating each of the RNN's own weights in differentiable fashion (internal storage) <ref name="J. Schmidhuber pages 191-195">{{Cite conference | author = [[Jürgen Schmidhuber]] | title = An introspective network that can learn to run its own weight change algorithm | booktitle= In Proc. of the Intl. Conf. on Artificial Neural Networks, Brighton | pages = 191–195 | publisher = IEE | year = 1993 | url = ftp://ftp.idsia.ch/pub/juergen/iee93self.ps.gz }}</ref><ref name="Hochreiter, Sepp 2001">{{cite journal | last1 = Hochreiter | first1 = Sepp | last2 = Younger | first2 = A. Steven | last3 = Conwell | first3 = Peter R. | year = 2001 | title = Learning to Learn Using Gradient Descent | url = | journal = ICANN | volume = 2130 | issue = | pages = 87–94 }}</ref>

*Learning to transduce with unbounded memory<ref name="Grefenstette, Edward 1506">Grefenstette, Edward, et al. [http://arxiv.org/pdf/1506.02516.pdf "Learning to Transduce with Unbounded Memory."] {{arxiv|1506.02516}} (2015).</ref>

Line 698:

====Memory networks====

Memory networks <ref name="Weston, Jason 1410">Weston, Jason, Sumit Chopra, and Antoine Bordes. "Memory networks." {{arxiv|1410.3916}} (2014).</ref><ref>Sukhbaatar, Sainbayar, et al. "End-To-End Memory Networks." {{arxiv|1503.08895}} (2015).</ref> is another extension to neural networks incorporating [[long-term memory]], which was developed by the [[Facebook]] research team. The long-term memory can be read and written to, with the goal of using it for prediction. These models have been applied in the context of [[question answering]] (QA) where the long-term memory effectively acts as a (dynamic) knowledge base, and the output is a textual response.<ref>Bordes, Antoine, et al. "Large-scale Simple Question Answering with Memory Networks." {{arxiv|1506.02075}} (2015).</ref>

====Pointer networks====

Deep neural networks can be potentially improved if they get deeper and have fewer parameters, while maintaining trainability. While training extremely deep (e.g. 1-million-layer-deep) neural networks might not be practically feasible, [[CPU]]-like architectures such as pointer networks <ref>Vinyals, Oriol, Meire Fortunato, and Navdeep Jaitly. "Pointer networks." {{arxiv|1506.03134}} (2015).</ref> and neural random-access machines <ref>Kurach,Karol, Andrychowicz, Marcin and Sutskever,Ilya. "Neural Random-Access Machines." {{arxiv|1511.06392}} (2015).</ref> developed by [[Google Brain]] researchers overcome this limitation by using external [[random-access memory]] as well as adding other components that typically belong to a [[computer architecture]] such as [[Processor register|registers]], [[Arithmetic logic unit|ALU]] and [[Pointer (computer programming)|pointers]]. Such systems operate on [[probability distribution]] vectors stored in memory cells and registers. Thus, the model is fully differentiable and trains end-to-end. The key characteristic of these models is that their depth, the size of their short-term memory, and the number of parameters can be altered independently — unlike models like [[LSTM]], whose number of parameters grows quadratically with memory size.

====Encoder–decoder networks====

Line 710:

=== Multilayer kernel machine ===

Multilayer kernel machines (MKM) as introduced in <ref name="ref59">{{cite journal

| last1 = Cho

| first1 = Youngmin

Line 771:

| Convolutional DNN w. Heterogeneous Pooling|| 18.7

|-

| Ensemble DNN/CNN/RNN <ref name=EnsembleDL>{{cite journal|last1=Deng|first1=L.|last2=Platt|first2=J.|title=Ensemble Deep Learning for Speech Recognition|journal=Proc. Interspeech|date=2014}}</ref>|| 18.2

|-

| Bidirectional LSTM|| 17.9

Line 794:

=== Image recognition ===

A common evaluation set for image classification is the [[MNIST database]] 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 allows multiple configurations to be tested. A comprehensive list of results on this set can be found in.<ref name="YANNMNIST">http://yann.lecun.com/exdb/mnist/.</ref> The current best result on MNIST is an error rate of 0.23%, achieved by Ciresan ''et al.'' in 2012.<ref name="CIRESAN2012">D. Ciresan, U. Meier, J. Schmidhuber., "Multi-column Deep Neural Networks for Image Classification," ''Technical Report No. IDSIA-04-12'', 2012.</ref>

According to LeCun,<ref name="lecun2016slides">[[Yann LeCun]] (2016). Slides on Deep Learning [https://indico.cern.ch/event/510372/ Online]</ref> in the early 2000s, in an industrial application CNNs already processed an estimated 10% to 20% of all the checks written in the US in the early 2000s.

Significant additional impact of deep learning in image or object recognition was felt in the years 2011-20122011–2012. Although CNNs trained by backpropagation had been around for decades<ref name="LECUN1989"/>, and GPU implementations of NNs for years,<ref name="jung2004"/> including CNNs,<ref name="chellapilla2006"/> fast implementations of CNNs with max-pooling on GPUs in the style of Dan Ciresan and colleagues<ref name="ciresan2011"/> were needed to make a dent in computer vision.<ref name="SCHIDHUB"/> In 2011, this approach achieved for the first time superhuman performance in a visual pattern recognition contest.<ref name="ciresan2011NN"/> Also in 2011, it won the ICDAR Chinese handwriting contest, and in May 2012, it won the ISBI image segmentation contest.<ref name="ciresan2012NIPS"/>

Until 2011, CNNs did not play a major role at computer vision conferences, but in June 2012, a paper by Dan Ciresan et al. at the leading conference CVPR<ref name="ciresan2011CVPR"/> showed how max-pooling CNNs on GPU can dramatically improve many vision benchmark records, sometimes with human-competitive or even superhuman performance. In October 2012, a similar system by Alex Krizhevsky in the team of Geoff Hinton<ref name="krizhevsky2012">Krizhevsky, A., Sutskever, I. and Hinton, G. E. (2012). ImageNet Classification with Deep Convolutional Neural Networks.

NIPS 2012: Neural Information Processing Systems, Lake Tahoe, Nevada</ref> won the large-scale ImageNet competition by a significant margin over shallow machine learning methods. In November 2012, Ciresan et al.'s system also won the ICPR contest on analysis of large medical images for cancer detection, and in the following year also the MICCAI Grand Challenge on the same topic<ref name="ciresan2013miccai">D. Ciresan, A. Giusti, L.M. Gambardella, J. Schmidhuber (2013). Mitosis Detection in Breast Cancer Histology Images using Deep Neural Networks. Proceedings MICCAI, 2013.</ref>

Line 808:

=== Natural language processing ===

Neural networks have been used for implementing [[language model]]s since the early 2000s.<ref name="gers2001"/><ref name="BENGIO2003">Y. Bengio, R. Ducharme, P. Vincent, C. Jauvin., "A Neural Probabilistic Language Model," ''Journal of Machine Learning Research'' 3 (2003) 1137–1155', 2003.</ref>

[[Recurrent neural networks]], especially LSTM,<ref name=lstm/> are most appropriate for sequential data such as language.

LSTM helped to improve machine translation<ref name="NIPS2014"/> and Language Modeling.<ref name="vinyals2016"/><ref name="gillick2015"/> LSTM combined with CNNs also improved automatic image captioning<ref name="vinyals2015"/> and a plethora of other applications.<ref name="SCHIDHUB"/>

Line 869:

== Theories of the human brain ==

Computational deep learning is closely related to a class of theories of [[brain development]] (specifically, neocortical development) proposed by [[cognitive neuroscientist]]s in the early 1990s.<ref name="UTGOFF">{{cite journal | last1 = Utgoff | first1 = P. E. | last2 = Stracuzzi | first2 = D. J. | year = 2002 | title = Many-layered learning | url = | journal = Neural Computation | volume = 14 | issue = | pages = 2497–2529 | doi=10.1162/08997660260293319}}</ref> An approachable summary of this work is Elman, et al.'s 1996 book "Rethinking Innateness" <ref name="ELMAN">J. Elman, ''et al.'', "Rethinking Innateness," 1996.</ref> (see also: Shrager and Johnson;<ref name="SHRAGER">{{cite journal | last1 = Shrager | first1 = J. | last2 = Johnson | first2 = MH | year = 1996 | title = Dynamic plasticity influences the emergence of function in a simple cortical array | url = | journal = Neural Networks | volume = 9 | issue = 7| pages = 1119–1129 | doi=10.1016/0893-6080(96)00033-0}}</ref> Quartz and Sejnowski <ref name="QUARTZ">{{cite journal | last1 = Quartz | first1 = SR | last2 = Sejnowski | first2 = TJ | year = 1997 | title = The neural basis of cognitive development: A constructivist manifesto | url = | journal = Behavioral and Brain Sciences | volume = 20 | issue = 4| pages = 537–556 | doi=10.1017/s0140525x97001581}}</ref>). As these developmental theories were also instantiated in computational models, they are technical predecessors of purely computationally motivated deep learning models. These developmental models share the interesting property that various proposed learning dynamics in the brain (e.g., a wave of [[nerve growth factor]]) conspire to support the [[self-organization]] of just the sort of inter-related neural networks utilized in the later, purely computational deep learning models; and such computational neural networks seem analogous to a view of the brain's neocortex as a hierarchy of filters in which each layer captures some of the information in the operating environment, and then passes the remainder, as well as modified base signal, to other layers further up the hierarchy. This process yields a self-organizing stack of [[transducer]]s, well-tuned to their operating environment. As described in The [[New York Times]] in 1995: "...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."<ref name="BLAKESLEE">S. Blakeslee., "In brain's early growth, timetable may be critical," ''The New York Times, Science Section'', pp. B5–B6, 1995.</ref>

The importance of deep learning with respect to the evolution and development of human [[cognition]] did not escape the attention of these researchers. One aspect of human development that distinguishes us from our nearest primate neighbors may be changes in the timing of development.<ref name="BUFILL">{BUFILL} E. Bufill, J. Agusti, R. Blesa., "Human neoteny revisited: The case of synaptic plasticity," ''American Journal of Human Biology'', 23 (6), pp. 729–739, 2011.</ref> Among [[primate]]s, the human brain remains relatively plastic until late in the post-natal period, whereas the brains of our closest relatives are more completely formed by birth. Thus, humans have greater access to the complex experiences afforded by being out in the world during the most formative period of brain development. This may enable us to "tune in" to rapidly changing features of the environment that other animals, more constrained by evolutionary structuring of their brains, are unable to take account of. To the extent that these changes are reflected in similar timing changes in hypothesized wave of cortical development, they may also lead to changes in the extraction of information from the stimulus environment during the early self-organization of the brain. Of course, along with this flexibility comes an extended period of immaturity, during which we are dependent upon our caretakers and our community for both support and training. The theory of deep learning therefore sees the coevolution of culture and cognition as a fundamental condition of human evolution.<ref name="SHRAGER2">J. Shrager and M. H. Johnson., "Timing in the development of cortical function: A computational approach," ''In B. Julesz and I. Kovacs (Eds.), Maturational windows and adult cortical plasticity'', 1995.</ref>

== Commercial activities ==

Deep learning is often presented as a step towards realising [[Artificial general intelligence|strong AI]]<ref name="HERN2013">D. Hernandez., "The Man Behind the Google Brain: Andrew Ng and the Quest for the New AI," ''http://www.wired.com/wiredenterprise/2013/05/neuro-artificial-intelligence/all/. Wired'', 10 May 2013.</ref> and thus many organizations have become interested in its use for particular applications. In December 2013, [[Facebook]] hired [[Yann LeCun]] to head its new [[artificial intelligence]] (AI) lab that was to have operations in California, London, and New York. The AI lab will develop deep learning techniques to help Facebook do tasks such as [[Automatic image annotation|automatically tagging uploaded pictures]] with the names of the people in them.<ref name="METZ2013">C. Metz., "Facebook's 'Deep Learning' Guru Reveals the Future of AI," ''http://www.wired.com/wiredenterprise/2013/12/facebook-yann-lecun-qa/. Wired'', 12 December 2013.</ref> Late in 2014, Facebook also hired [[Vladimir Vapnik]], a main developer of the Vapnik–Chervonenkis theory of statistical learning, and co-inventor of the [[support vector machine]] method.<ref name="Vapnik2014">V. Vapnik., "research.facebook.com" ''[https://research.facebook.com/researchers/1566384816909948/vladimir-vapnik/'', "research.facebook.com"] .</ref>

In March 2013, [[Google]] hired [[Geoffrey Hinton]] and two of his graduate students, Alex Krizhevsky and Ilya Sutskever. Their work was to focus on both improving existing machine learning products at Google and help deal with the growing amount of data Google has. Google also bought Hinton's company, DNNresearch.

Line 892:

A main criticism of deep learning concerns the lack of theory surrounding many of the methods. Most of the learning in deep architectures is just some form of [[gradient descent]]. While gradient descent has been understood for a while now, the theory surrounding other algorithms, such as contrastive divergence is less clear (i.e., Does it converge? If so, how fast? What is it approximating?). Deep learning methods are often looked at as a black box, with most confirmations done empirically, rather than theoretically.

Others point out that deep learning should be looked at as a step towards realizing strong AI, not as an all-encompassing solution. Despite the power of deep learning methods, they still lack much of the functionality needed for realizing this goal entirely. Research psychologist [[Gary Marcus]] has noted that:<blockquote>"Realistically, deep learning is only part of the larger challenge of building intelligent machines. Such techniques lack ways of representing [[causality|causal relationships]] (...) have no obvious ways of performing [[inference|logical inferences]], 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 [[Watson (computer)|Watson]] (...) use techniques like deep learning as just one element in a very complicated ensemble of techniques, ranging from the statistical technique of [[Bayesian inference]] to [[deductive reasoning]]." <ref name="MARCUS">G. Marcus., "Is "Deep Learning" a Revolution in Artificial Intelligence?" ''The New Yorker'', 25 November 2012.</ref></blockquote>To the extent that such a viewpoint implies, without intending to, that deep learning will ultimately constitute nothing more than the primitive discriminatory levels of a comprehensive future machine intelligence, a recent pair of speculations regarding art and artificial intelligence<ref>{{cite web |url=http://artent.net/2015/03/27/art-and-artificial-intelligence-by-g-w-smith/|author=Smith, G. W.|date=March 27, 2015|title=Art and Artificial Intelligence|publisher=ArtEnt|accessdate=March 27, 2015}}</ref> offers an alternative and more expansive outlook. The first such speculation is that it might be possible to train a machine vision stack to perform the sophisticated task of discriminating between "old master" and amateur figure drawings; and the second is that such a sensitivity might in fact represent the rudiments of a non-trivial machine empathy. It is suggested, moreover, that such an eventuality would be in line with anthropology, which identifies a concern with aesthetics as a key element of [[behavioral modernity]] (eg, in <ref>{{cite web |url=https://www.newscientist.com/article/dn1528-tools-point-to-african-origin-for-human-behaviour/ |author=Knight, Will |date=November 7, 2001 |title=Tools point to African origin for human behaviour |publisher=New Scientist |accessdate=October 7, 2015 }}</ref>).

In further reference to the idea that a significant degree of artistic sensitivity might inhere within relatively low levels, whether biological or digital, 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 on which they were trained<ref>{{cite web |url=http://googleresearch.blogspot.co.uk/2015/06/inceptionism-going-deeper-into-neural.html|author=Alexander Mordvintsev, Christopher Olah, and Mike Tyka|date=June 17, 2015|title=Inceptionism: Going Deeper into Neural Networks|publisher=Google Research Blog|accessdate=June 20, 2015}}</ref> seem to demonstrate a striking visual appeal in light of the remarkable level of public attention which this work captured: the original research notice received well over 1,000 comments, and the [[The Guardian]] coverage<ref>{{cite web |url=http://www.theguardian.com/technology/2015/jun/18/google-image-recognition-neural-network-androids-dream-electric-sheep|author=Alex Hern|date=June 18, 2015|title=Yes, androids do dream of electric sheep|publisher=The Guardian|accessdate=June 20, 2015}}</ref> was for a time the most frequently accessed article on that newspaper's web site.

Some currently popular and successful deep learning architectures display certain problematic behaviors,<ref name=goertzel>Ben Goertzel. Are there Deep Reasons Underlying the Pathologies of Today's Deep Learning Algorithms? (2015) Url: http://goertzel.org/DeepLearning_v1.pdf</ref> such as confidently classifying random data as belonging to a familiar category of nonrandom images<ref>Nguyen, Anh, Jason Yosinski, and Jeff Clune. "Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Images." {{arxiv|1412.1897}} (2014).</ref> and misclassifying minuscule perturbations of correctly classified images.<ref>Szegedy, Christian, et al. "Intriguing properties of neural networks." {{arxiv|1312.6199}} (2013).</ref> The creator of [[OpenCog]], [[Ben Goertzel]], hypothesized<ref name="goertzel"/> that these behaviors are due to limitations in the internal representations learned by these architectures, and that these limitations would inhibit integration of these architectures into heterogeneous multi-component [[Artificial general intelligence|AGI]] architectures. It is suggested that these issues can be worked around by developing deep learning architectures that internally form states homologous to image-grammar<ref>{{cite journal | last1 = Zhu | first1 = S.C. | last2 = Mumford | first2 = D. | year = | title = A stochastic grammar of images | url = | journal = Found. Trends Comput. Graph. Vis. | volume = 2 | issue = 4| pages = 259–362 | doi = 10.1561/0600000018 }}</ref> decompositions of observed entities and events.<ref name="goertzel"/> Learning a [[grammar]] (visual or linguistic) from training data would be equivalent to restricting the system to [[commonsense reasoning]] which operates on concepts in terms of [[Production (computer science)|production rules]] of the grammar, and is a basic goal of both human language acquisition <ref>Miller, G. A., and N. Chomsky. "Pattern conception." Paper for Conference on pattern detection, University of Michigan. 1957.</ref> and AI. (See also [[Grammar induction]].<ref>Jason Eisner, Deep Learning of Recursive Structure: Grammar Induction, http://techtalks.tv/talks/deep-learning-of-recursive-structure-grammar-induction/58089/</ref>)

== Software libraries ==