Generalization in Interactive Networks: The Benefits of Inhibitory Competition and Hebbian Learning
The DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Abstract
Computational models in cognitive neuroscience should ideally use biological properties and powerful computational principles to produce behavior consistent with psychological findings. Error-driven backpropagation is computationally powerful and has proven useful for modeling a range of psychological data but is not biologically plausible. Several approaches to implementing backpropagation in a biologically plausible fashion converge on the idea of using bid irectional activation propagation in interactive networks to convey error signals. This article demonstrates two main points about these error-driven interactive networks: (1) they generalize poorly due to attractor dynamics that interfere with the network's ability to produce novel combinatorial representations systematically in response to novel inputs, and (2) this generalization problem can be remedied by adding two widely used mechanistic principles, inhibitory competition and Hebbian learning, that can be independently motivated for a variety of biological, psychological, and computational reasons. Simulations using the Leabra algorithm, which combines the generalized recirculation (GeneRec), biologically plausible, error-driven learning algorithm with inhibitory competition and Hebbian learning, show that these mechanisms can result in good generalization in interactive networks. These results support the general conclusion that cognitive neuroscience models that incorporate the core mechanistic principles of interactivity, inhibitory competition, and error-driven and Hebbian learning satisfy a wider range of biological, psychological, and computational constraints than models employing a subset of these principles.
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
Neural networks and physical systems with emergent collective computational abilities.
Hopfield, J. J. (1982) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.79.8.2554 · Cited by 19,270 · cited 1×
“Computa-tionally, interactivity enables iterative, constraint satisfaction style processing (Smolensky, 1986; Ackley, Hinton, & Sejnowski, 1985; Hopfield, 1982, 1984), and generally allows information processing to flow in many different directions within one network, providing a richer and more…”
BackgroundSelf-Organization and Associative Memory
Kohonen, T. (1989) · Springer Series in Information Sciences · doi:10.1007/978-3-642-88163-3 · Cited by 8,740
Hopfield, J. J. (1984) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.81.10.3088 · Cited by 6,918
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×
“To further substantiate the argument in favor of using sparse distributed representations, Olshausen and Field (1996) developed a simulation that showed that imposing a bias for developing this type of representation can result in the development of realistic early visual representations (oriented…”
MethodMcClelland, J. L., McNaughton, B. L., & O'Reilly, R. C. (1995) · Psychological Review · doi:10.1037/0033-295x.102.3.419 · Cited by 5,256 · cited 2×
“Glushko (1979) regulars 95.3 97.7 93.8 Glushko (1979) exceptions 97.6 100.0 95.9 McCann & Besner (1987) controls 85.9 85.0 88.6 McCann & Besner (1987) homophones 92.3 N.A.”
InfluentialSelf-citationMcClelland, J. L., & Rumelhart, D. E. (1981) · Psychological Review · doi:10.1037/0033-295x.88.5.375 · Cited by 4,889
A 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
Neural Networks and the Bias/Variance Dilemma
Geman, S., Bienenstock, E., & Doursat, R. (1992) · Neural Computation · doi:10.1162/neco.1992.4.1.1 · Cited by 3,578 · cited 2×
“It is well-known that purely error-driven learning mechanisms are weakly biased, and are therefore typically underconstrained by the learning task, causing them to suffer from too much variance (e.g., Vapnik & Chervonenkis, 1971; Weigend et al., 1991; Geman et al., 1992; Wolpert, 1992).”
BackgroundA 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
Simple memory: a theory for archicortex
Marr, D. (1971) · Philosophical Transactions of the Royal Society of London. B, Biological Sciences · doi:10.1098/rstb.1971.0078 · Cited by 2,771 · cited 1×
“…in the context of hippocampal representations, which are thought to be conjunctive, while cortical representations are thought to be more elemental or independent (e.g., Marr, 1971; O’Reilly & McClelland, 1994; McClelland, McNaughton, & O’Reilly, 1995; O’Reilly & Rudy, in press-b, in press-a).”
BackgroundUnderstanding 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 7×
“…model presented in O’Reilly and Munakata (2000) replicated the generalization performance of the backprop networks reported in Plaut et al. (1996), but did so using simpler and more surface-valid input/output representations than the carefully-crafted representations used by Plaut et al. (1996).”
InfluentialOn the Uniform Convergence of Relative Frequencies of Events to Their Probabilities
Vapnik, V. N., & Chervonenkis, A. Y. (1971) · Theory of Probability & Its Applications · doi:10.1137/1116025 · Cited by 2,448
Simplified neuron model as a principal component analyzer
Oja, E. (1982) · Journal of Mathematical Biology · doi:10.1007/bf00275687 · Cited by 2,309 · cited 2×
“Leabra uses essentially the same learning rule used in competitive learning or mixtures-of-gaussians (Rumelhart & Zipser, 1986; Nowlan, 1990), which can be seen as a variant of the Oja normalization (Oja, 1982).”
MethodSynaptic plasticity: LTP and LTD
Bear, M. F., & Malenka, R. C. (1994) · Current Opinion in Neurobiology · doi:10.1016/0959-4388(94)90101-5 · Cited by 1,355
A mechanism for the Hebb and the anti-Hebb processes underlying learning and memory.
Lisman, J. (1989) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.86.23.9574 · Cited by 1,127 · cited 1×
“Hebbian synaptic modification appears to operate on the concentration of postsynaptic calcium ions — lower levels of calcium produce LTD, and higher levels produce LTP (Artola, Brocher, & Singer, 1990; Lisman, 1989, 1994; Bear & Malenka, 1994).”
BackgroundThe organization and activation of orthographic knowledge in reading aloud.
Glushko, R. J. (1979) · Journal of Experimental Psychology: Human Perception and Performance · doi:10.1037//0096-1523.5.4.674 · Cited by 961
Generalization of back-propagation to recurrent neural networks
Pineda, F. J. (1987) · Physical Review Letters · doi:10.1103/physrevlett.59.2229 · Cited by 931 · cited 2×
“dress this concern and provide additional evidence for the idea that interactivity impairs generalization, networks using the Almeida-Pineda (AP) exed-point recurrent backpropagation algorithm (Almeida, 1987; Pineda, 1987a, 1987b, 1988 ) were also run.”
MethodHippocampal conjunctive encoding, storage, and recall: Avoiding a trade‐off
O'Reilly, R. C., & McClelland, J. L. (1994) · Hippocampus · doi:10.1002/hipo.450040605 · Cited by 914
Self-citationConjunctive representations in learning and memory: Principles of cortical and hippocampal function.
O'Reilly, R. C., & Rudy, J. W. (2001) · Psychological Review · doi:10.1037//0033-295x.108.2.311 · Cited by 865
Self-citationO'Reilly, R. C. (1996) · Neural Computation · doi:10.1162/neco.1996.8.5.895 · Cited by 352
Self-citationConspiracy effects in word pronunciation
Taraban, R., & McClelland, J. L. (1987) · Journal of Memory and Language · doi:10.1016/0749-596x(87)90105-7 · Cited by 301
McCann, R. S., & Besner, D. (1987) · Journal of Experimental Psychology: Human Perception and Performance · doi:10.1037/0096-1523.13.1.14 · Cited by 262
Cited by
O'Reilly, R. C., & Frank, M. J. (2006) · Neural Computation · doi:10.1162/089976606775093909 · Cited by 1,066
Self-citationConjunctive representations in learning and memory: Principles of cortical and hippocampal function.
O'Reilly, R. C., & Rudy, J. W. (2001) · Psychological Review · doi:10.1037//0033-295x.108.2.311 · Cited by 865
Self-citationPerry, C., Ziegler, J. C., & Zorzi, M. (2007) · Psychological Review · doi:10.1037/0033-295x.114.2.273 · Cited by 748
Hold your horses: a dynamic computational role for the subthalamic nucleus in decision making
Frank, M. J. (2006) · Neural Networks · doi:10.1016/j.neunet.2006.03.006 · Cited by 706
“The model is implemented using the Leabra framework (O’Reilly, 2001; O’Reilly & Munakata, 2000).”
Karbach, J., & Verhaeghen, P. (2014) · Psychological Science · doi:10.1177/0956797614548725 · Cited by 589
Metadata
- Authors
- O'Reilly, R. C.
- Venue
- Neural Computation
- Publisher
- MIT Press - Journals
- Published
- 2001-06-01
- Volume
- 13
- Issue
- 6
- Pages
- 1199-1241
- ISSN
- 0899-7667 (ISSN-L)
- 1530-888X
- OpenAlex
- W2153758134
- Cited by
- 132
- References
- 22
- FWCI
- 7.51
- Type
- journal article
- Language
- en
- Open access
- closed
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.