Modeling language and cognition with deep unsupervised learning: a tutorial overview
Open accessThe DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index.
Abstract
Deep unsupervised learning in stochastic recurrent neural networks with many layers of hidden units is a recent breakthrough in neural computation research. These networks build a hierarchy of progressively more complex distributed representations of the sensory data by fitting a hierarchical generative model. In this article we discuss the theoretical foundations of this approach and we review key issues related to training, testing and analysis of deep networks for modeling language and cognitive processing. The classic letter and word perception problem of McClelland and Rumelhart (1981) is used as a tutorial example to illustrate how structured and abstract representations may emerge from deep generative learning. We argue that the focus on deep architectures and generative (rather than discriminative) learning represents a crucial step forward for the connectionist modeling enterprise, because it offers a more plausible model of cortical learning as well as a way to bridge the gap between emergentist connectionist models and structured Bayesian models of cognition.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805
Learning representations by back-propagating errors
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986) · Nature · doi:10.1038/323533a0 · Cited by 31,892
The perceptron: A probabilistic model for information storage and organization in the brain.
Rosenblatt, F. (1958) · Psychological Review · doi:10.1037/h0042519 · Cited by 12,280
Kohonen, T. (1990) · Proceedings of the IEEE · doi:10.1109/5.58325 · Cited by 8,414
Whatever next? Predictive brains, situated agents, and the future of cognitive science
Clark, A. (2013) · Behavioral and Brain Sciences · doi:10.1017/s0140525x12000477 · Cited by 5,924
Emergence of simple-cell receptive field properties by learning a sparse code for natural images
Olshausen, B. A., & Field, D. J. (1996) · Nature · doi:10.1038/381607a0 · Cited by 5,694 · cited 1×
“In this case the network develops sparse distributed representations, which have many useful properties and appear to be a coding strategy adopted by the brain (Olshausen and Field, 1996; see Olshausen and Field, 2004, for review).”
MethodRao, R. P. N., & Ballard, D. H. (1999) · Nature Neuroscience · doi:10.1038/4580 · Cited by 5,483 · cited 1×
“Notably, the application of these algorithms to natural images has been shown to generate receptive field properties similar to those observed in the visual cortex (Rao and Ballard, 1999).”
Training Products of Experts by Minimizing Contrastive Divergence
Hinton, G. E. (2002) · Neural Computation · doi:10.1162/089976602760128018 · Cited by 4,879 · cited 6×
“Importantly, a restriction to the architecture of the BM by not allowing intra-layer connections (RBM; Hinton, 2002) makes learningextremely fast.”
BackgroundMethodInfluentialA theory of cortical responses.
Friston, K. (2005) · Philosophical Transactions of the Royal Society B: Biological Sciences · doi:10.1098/rstb.2005.1622 · Cited by 4,578 · cited 1×
“Notably, the learned internal model can then be used to infer causes and make predictions (Dayan et al., 1995; Hinton and Ghahramani, 1997; Friston, 2005; Hinton, 2010b; Huang and Rao, 2011; Clark, 2013).”
MethodA distributed, developmental model of word recognition and naming.
Seidenberg, M. S., & McClelland, J. L. (1989) · Psychological Review · doi:10.1037/0033-295x.96.4.523 · Cited by 3,909
A learning algorithm for Boltzmann machines
Ackley, D. H., Hinton, G. E., & Sejnowski, T. J. (1985) · Cognitive Science · doi:10.1207/s15516709cog0901_7 · Cited by 3,402
Hierarchical models of object recognition in cortex
Riesenhuber, M., & Poggio, T. (1999) · Nature Neuroscience · doi:10.1038/14819 · Cited by 3,309 · cited 1×
“Hierarchical processing is thought to be a fundamental characteristic of cortical computation (Hinton, 2007; Clark, 2013) and it is a key feature of biologically inspired computational models of vision (Riesenhuber and Poggio, 1999).”
BackgroundA Practical Guide to Training Restricted Boltzmann Machines
Hinton, G. E. (2012) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-35289-8_32 · Cited by 2,852
Understanding normal and impaired word reading: Computational principles in quasi-regular domains.
Plaut, D. C., McClelland, J. L., Seidenberg, M. S., & Patterson, K. (1996) · Psychological Review · doi:10.1037/0033-295x.103.1.56 · Cited by 2,698 · cited 2×
“…include reaction times and accuracy of skilled readers at the level of individual words, the development of reading skills in children, and the impaired performance of dyslexic individuals (Plaut et al., 1996; Zorzi et al., 1998; Harm and Seidenberg, 1999, 2004; Perry et al., 2007, 2010, 2013).”
An introduction to MCMC for machine learning.
Andrieu, C., de Freitas, N., Doucet, A., & Jordan, M. I. (2003) · Machine Learning · doi:10.1023/a:1020281327116 · Cited by 2,451
Lee, H., Grosse, R., Ranganath, R., & Ng, A. Y. (2009) · Proceedings of the 26th Annual International Conference on Machine Learning · doi:10.1145/1553374.1553453 · Cited by 2,412
Simplified neuron model as a principal component analyzer
Oja, E. (1982) · Journal of Mathematical Biology · doi:10.1007/bf00275687 · Cited by 2,309
A View of the Em Algorithm that Justifies Incremental, Sparse, and other Variants
Neal, R. M., & Hinton, G. E. (1998) · Learning in Graphical Models · doi:10.1007/978-94-011-5014-9_12 · Cited by 2,187 · cited 1×
“A good compromise between these two approaches is to use a mini-batch learning scheme, in which the dataset is partitioned into small subsets (i.e., mini-batches) and the weights are updated with the average gradient computed on each subset (Neal and Hinton, 1998).”
MethodHow does the brain solve visual object recognition?
DiCarlo, J. J., Zoccolan, D., & Rust, N. C. (2012) · Neuron · doi:10.1016/j.neuron.2012.01.010 · Cited by 1,790 · cited 1×
“In the case of real sensory signals, this shortcoming is exacerbated by the fact that the feature vectors are high-dimensional and usually lie on highly curved and tangled manifolds (DiCarlo et al., 2012).”
BackgroundBuilding high-level features using large scale unsupervised learning
Le, Q. V. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6639343 · Cited by 1,695
Sparse coding of sensory inputs
Olshausen, B. A., & Field, D. J. (2004) · Current Opinion in Neurobiology · doi:10.1016/j.conb.2004.07.007 · Cited by 1,404
Perry, C., Ziegler, J. C., & Zorzi, M. (2007) · Psychological Review · doi:10.1037/0033-295x.114.2.273 · Cited by 748
Self-citationLarge-scale deep unsupervised learning using graphics processors
Raina, R., Madhavan, A., & Ng, A. Y. (2009) · Proceedings of the 26th Annual International Conference on Machine Learning · doi:10.1145/1553374.1553486 · Cited by 742 · cited 1×
“If implemented on multicore hardware, deep learning is practical even with billions of connections, thereby allowing the development of very-large-scale simulations (Raina et al., 2009; Dean et al., 2012; Le et al., 2012).”
BackgroundVinckier, F., Dehaene, S., Jobert, A., Dubus, J. P., Sigman, M., & Cohen, L. (2007) · Neuron · doi:10.1016/j.neuron.2007.05.031 · Cited by 717 · cited 1×
“The idea that perception of written words involves the sensitivity to increasingly larger orthographic units is also supported by recent neuroimaging findings (Dehaene et al., 2005; Vinckier et al., 2007).”
ResultProbabilistic models of cognition: exploring representations and inductive biases
Griffiths, T. L., Chater, N., Kemp, C., Perfors, A., & Tenenbaum, J. B. (2010) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2010.05.004 · Cited by 568
Huang, Y., & Rao, R. P. N. (2011) · WIREs Cognitive Science · doi:10.1002/wcs.142 · Cited by 470
Perry, C., Ziegler, J. C., & Zorzi, M. (2010) · Cognitive Psychology · doi:10.1016/j.cogpsych.2010.04.001 · Cited by 413
Self-citationProbabilistic models of language processing and acquisition
Chater, N., & Manning, C. D. (2006) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2006.05.006 · Cited by 411 · cited 2×
“Structured Bayesian models of cognition (for reviews see Chater et al., 2006; Griffiths et al., 2010) assume that human learning and inference approximately follow the principles of Bayesian probabilistic inference and they have been used in the last few years to address a number of issues in…”
MethodTwo routes or one in reading aloud? A connectionist dual-process model.
Zorzi, M., Houghton, G., & Butterworth, B. (1998) · Journal of Experimental Psychology: Human Perception and Performance · doi:10.1037/0096-1523.24.4.1131 · Cited by 339
Self-citationEmergence of a 'visual number sense' in hierarchical generative models
Stoianov, I., & Zorzi, M. (2012) · Nature Neuroscience · doi:10.1038/nn.2996 · Cited by 300 · cited 8×
“This aspect was present in two additional problems that complemented our tutorial with more realistic simulations, that is, handwritten digit recognition (LeCun et al., 1998) and visual numerosity perception (Stoianov and Zorzi, 2012).”
InfluentialSelf-citationInteractive Activation and Mutual Constraint Satisfaction in Perception and Cognition
McClelland, J. L., Mirman, D., Bolger, D. J., & Khaitan, P. (2014) · Cognitive Science · doi:10.1111/cogs.12146 · Cited by 122
Cited by
Mandera, P., Keuleers, E., & Brysbaert, M. (2017) · Journal of Memory and Language · doi:10.1016/j.jml.2016.04.001 · Cited by 428
Deep temporal models and active inference
Friston, K. J., Rosch, R., Parr, T., Price, C., & Bowman, H. (2017) · Neuroscience & Biobehavioral Reviews · doi:10.1016/j.neubiorev.2017.04.009 · Cited by 237
“This form of modelling has been important in machine learning; e.g. (Ne fi an et al., 2002; George and Hawkins, 2009), with a special focus on hierarchical or deep architectures (Salakhutdinov et al., 2013; Zorzi et al., 2013; Testolin and Zorzi, 2016).”
Deep temporal models and active inference
Friston, K. J., Rosch, R., Parr, T., Price, C., & Bowman, H. (2018) · Neuroscience & Biobehavioral Reviews · doi:10.1016/j.neubiorev.2018.04.004 · Cited by 119
“, 2002; George and Hawkins, 2009), with a special focus on hierarchical or deep architectures (Salakhutdinov et al., 2013; Zorzi et al., 2013; Testolin and Zorzi, 2016).”
BackgroundTestolin, A., Zanforlin, M., De Filippo De Grazia, M., Munaretto, D., Zanella, A., Zorzi, M., & Zorzi, M. (2014) · 2014 13th Annual Mediterranean Ad Hoc Networking Workshop (MED-HOC-NET) · doi:10.1109/medhocnet.2014.6849102 · Cited by 45
“We use an equivalent but computationally more efficient method, which relies on the calculation of a pseudo-inverse matrix and is readily available in some high-level programming languages such as Python or MATLAB [22].”
BackgroundMethodInfluentialSelf-citationFederated inference and belief sharing
Friston, K. J., Parr, T., Heins, C., Constant, A., Friedman, D., Isomura, T., Fields, C., Verbelen, T., Ramstead, M., Clippinger, J., et al. (2024) · Neuroscience & Biobehavioral Reviews · doi:10.1016/j.neubiorev.2023.105500 · Cited by 30
“See for example: (Davis and Johnsrude, 2003; Friston et al., 2017c; George and Hawkins, 2009; MacKay and Peto, 2008; Stoianov et al., 2022; Yildiz et al., 2013; Young et al., 2018; Zorzi et al., 2013).”
Metadata
- Authors
- Zorzi, M., Testolin, A., & Stoianov, I. P.
- Venue
- Frontiers in Psychology
- Publisher
- Frontiers Media SA
- Published
- 2013
- Volume
- 4
- Pages
- 515-515
- ISSN
- 1664-1078 (ISSN-L)
- OpenAlex
- W2124151298
- Cited by
- 67
- References
- 32
- FWCI
- 8.47
- Type
- journal article
- Language
- en
- Open access
- gold
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.