Prediction-error driven learning: the engine of change in cognitive development

Summary form only given. Since the introduction of powerful connectionist learning procedures such as back propagation, our group has been studying the use of such procedures as a mechanism for cognitive development. The models can be viewed as capturing Piaget's idea that development is essentially an experience driven process, dependent of the adjustment of schemas to assimilate disparities between expectations and observations. Among the issues that have been addressed with these models are the following: (1) Is an initial naive domain theory required to allow a learner to interpret experience in a given domain? (2) Why does learning appear to occur in stages, punctuated by plateaus? (3) what does it mean to be 'ready to learn'? (4) Is it necessary to start small in learning, and if so when and why? (5) Why are there critical or sensitive periods in learning; why is it sometimes easier to learn when a system is 'young' or inexperienced?.

[1]  J. Gibbon.  Scalar expectancy theory and Weber's law in animal timing. , 1977 .

[2]  R. Morris.  Parallel Distributed Processing: Implications for Psychology and Neurobiology , 1990 .

[3]  R. Rescorla.  A theory of pavlovian conditioning: The effectiveness of reinforcement and non-reinforcement , 1972 .

[4]  Bernard Widrow,et al.  Perceptrons, adalines, and backpropagation , 1998 .

[5]  David S. Touretzky,et al.  Operant Conditioning in Skinnerbots , 1997, Adapt. Behav..

[6]  James L. McClelland,et al.  Graded state machines: The representation of temporal contingencies in simple recurrent networks , 1991, Machine Learning.

[7]  J. Pearce.  Similarity and discrimination: a selective review and a connectionist model. , 1994, Psychological review.

[8]  J. Flavell.  The Developmental psychology of Jean Piaget , 1963 .

[9]  Peter Dayan,et al.  A Neural Substrate of Prediction and Reward , 1997, Science.

[10]  Benjamin Van Roy,et al.  Average cost temporal-difference learning , 1997, Proceedings of the 36th IEEE Conference on Decision and Control.

[11]  James L. McClelland,et al.  Rethinking infant knowledge: toward an adaptive process account of successes and failures in object permanence tasks. , 1997, Psychological review.

[12]  Joel L. Davis,et al.  A Model of How the Basal Ganglia Generate and Use Neural Signals That Predict Reinforcement , 1994 .

[13]  James L. McClelland,et al.  On learning the past-tenses of English verbs: implicit rules or parallel distributed processing , 1986 .

[14]  A. Kacelnik.  Normative and descriptive models of decision making: time discounting and risk sensitivity. , 2007, Ciba Foundation symposium.

[15]  P. Dayan,et al.  A framework for mesencephalic dopamine systems based on predictive Hebbian learning , 1996, The Journal of neuroscience : the official journal of the Society for Neuroscience.

[16]  Stephen Grossberg,et al.  Neural dynamics of adaptive timing and temporal discrimination during associative learning , 1989, Neural Networks.

[17]  David S. Touretzky,et al.  Behavioral considerations suggest an average reward TD model of the dopamine system , 2000, Neurocomputing.

[18]  James L. McClelland,et al.  Understanding normal and impaired word reading: computational principles in quasi-regular domains. , 1996, Psychological review.

[19]  JOHN W. Moore,et al.  To appear in D.A. Rosenbaum & C.E. Collyer (Eds.), Timing of behavior: Neural, computational, and psychological perspectives. Cambridge, MA: MIT Press Predictive Timing Under Temporal Uncertainty: The TD Model of the Conditioned Response , 1996 .

[20]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.