Towards a more flexible language of thought: Bayesian grammar updates after each concept exposure.

Recent approaches to human concept learning have successfully combined the power of symbolic, infinitely productive rule systems and statistical learning to explain our ability to learn new concepts from just a few examples. The aim of most of these studies is to reveal the underlying language structuring these representations and providing a general substrate for thought. However, describing a model of thought that is fixed once trained is against the extensive literature that shows how experience shapes concept learning. Here, we ask about the plasticity of these symbolic descriptive languages. We perform a concept learning experiment that demonstrates that humans can change very rapidly the repertoire of symbols they use to identify concepts, by compiling expressions that are frequently used into new symbols of the language. The pattern of concept learning times is accurately described by a Bayesian agent that rationally updates the probability of compiling a new expression according to how useful it has been to compress concepts so far. By portraying the language of thought as a flexible system of rules, we also highlight the difficulties to pin it down empirically.

[1]  N. Chater,et al.  Simplicity: a unifying principle in cognitive science? , 2003, Trends in Cognitive Sciences.

[2]  Michael I. Jordan,et al.  Learning Programs: A Hierarchical Bayesian Approach , 2010, ICML.

[3]  E. Leeuwenberg.  A perceptual coding language for visual and auditory patterns. , 1971, The American journal of psychology.

[4]  Thomas L. Griffiths,et al.  A Rational Analysis of Rule-Based Concept Learning , 2008, Cogn. Sci..

[5]  John K. Kruschke,et al.  Dimensional Relevance Shifts in Category Learning , 1996, Connect. Sci..

[6]  J. Fodor,et al.  The Language of Thought , 1980 .

[7]  Mariano Sigman,et al.  The language of geometry: Fast comprehension of geometrical primitives and rules in human adults and preschoolers , 2017, PLoS Comput. Biol..

[8]  Peter M. Vishton,et al.  Rule learning by seven-month-old infants. , 1999, Science.

[9]  Noah D. Goodman,et al.  The logical primitives of thought: Empirical foundations for compositional cognitive models. , 2016, Psychological review.

[10]  A. Kolmogorov.  Three approaches to the quantitative definition of information , 1968 .

[11]  Gary Lupyan,et al.  Language is not Just for Talking , 2007, Psychological science.

[12]  D. Gentner,et al.  Language in Mind: Advances in the Study of Language and Thought , 2003 .

[13]  Noah D. Goodman,et al.  Bootstrapping in a language of thought: A formal model of numerical concept learning , 2012, Cognition.

[14]  Pascal Vincent,et al.  Representation Learning: A Review and New Perspectives , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[15]  D. Everett.  Cultural Constraints on Grammar and Cognition in Pirahã , 2005, Current Anthropology.

[16]  S. Gelman.  Learning from others: children's construction of concepts. , 2009, Annual review of psychology.

[17]  R. Mathews,et al.  Role of Implicit and Explicit Processes in Learning From Examples: A Synergistic Effect , 1989 .

[18]  Marvin Minsky,et al.  Computation : finite and infinite machines , 2016 .

[19]  John R. Anderson.  ACT: A simple theory of complex cognition. , 1996 .

[20]  Gregory L. Murphy,et al.  Comprehending Complex Concepts , 1988, Cogn. Sci..

[21]  Noah D. Goodman,et al.  Theory Acquisition and the Language of Thought , 2008 .

[22]  A. Reber.  Implicit learning and tacit knowledge , 1993 .

[23]  Ray J. Solomonoff,et al.  A Formal Theory of Inductive Inference. Part II , 1964, Inf. Control..

[24]  J. Rissanen,et al.  Modeling By Shortest Data Description* , 1978, Autom..

[25]  S. Levinson,et al.  Opinion TRENDS in Cognitive Sciences Vol.8 No.3 March 2004 Can language restructure cognition? The case for space , 2022 .

[26]  Noam Chomsky.  Knowledge of language: its nature, origin, and use , 1988 .

[27]  Ronaldo Vigo,et al.  A note on the complexity of Boolean concepts , 2006 .

[28]  Allen Newell,et al.  Chunking in Soar: The anatomy of a general learning mechanism , 1985, Machine Learning.

[29]  J. Feldman.  The Simplicity Principle in Human Concept Learning , 2003 .

[30]  Oren Etzioni,et al.  Explanation-Based Learning: A Problem Solving Perspective , 1989, Artif. Intell..

[31]  Robert L. Goldstone,et al.  The development of features in object concepts , 1998, Behavioral and Brain Sciences.

[32]  Mariano Sigman,et al.  Bayesian validation of grammar productions for the language of thought , 2018, PloS one.

[33]  Allen Newell,et al.  Physical Symbol Systems , 1980, Cogn. Sci..

[34]  Robert L. Goldstone,et al.  Putting category learning in order: Category structure and temporal arrangement affect the benefit of interleaved over blocked study , 2014, Memory & cognition.

[35]  S. Denison,et al.  Rational variability in children’s causal inferences: The Sampling Hypothesis , 2013, Cognition.

[36]  Charles Kemp,et al.  Exploring the conceptual universe. , 2012, Psychological review.

[37]  Joshua B. Tenenbaum,et al.  Human-level concept learning through probabilistic program induction , 2015, Science.

[38]  Samuel J. Gershman,et al.  Computational rationality: A converging paradigm for intelligence in brains, minds, and machines , 2015, Science.

[39]  Marcus R. Watson,et al.  Extremely selective attention: eye-tracking studies of the dynamic allocation of attention to stimulus features in categorization. , 2009, Journal of experimental psychology. Learning, memory, and cognition.