Traditional Perceptrons Do Not Produce the Overexpectation Effect

Perceptrons are typically viewed as being an artificial neural network that embodies the Rescorla-Wagner model of learning. One of the important properties of the Rescorla-Wagner model was its prediction of the overexpectation effect. However, we show below that a typical perceptron is not capable of generating this effect. This result brings into question assumed relationships between artificial neural networks and models of animal learning.

[1]  M. Dawson,et al.  Minds and Machines: Connectionism and Psychological Modeling , 2003 .

[2]  M. Escobar,et al.  Recovery from the overexpectation effect : Contrasting performance-focused and acquisition-focused models of retrospective revaluation , 2002 .

[3]  Blocking, unblocking, and overexpectation in autoshaping with pigeons. , 1996, Journal of the experimental analysis of behavior.

[4]  David R. Shanks,et al.  The Psychology of Associative Learning , 1995 .

[5]  Sadahiko Nakajima,et al.  Overexpectation in appetitive Pavlovian and instrumental conditioning , 1998 .

[6]  J. Kruschke.  Attention in Learning , 2003 .

[7]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[8]  E. Kremer.  The Rescorla-Wagner model: losses in associative strength in compound conditioned stimuli. , 1978, Journal of experimental psychology. Animal behavior processes.

[9]  N. Mackintosh.  A Theory of Attention: Variations in the Associability of Stimuli with Reinforcement , 1975 .

[10]  J. Houston.  Generalization and discrimination , 1981 .

[11]  Gavan P McNally,et al.  Blocking, unblocking, and overexpectation of fear: a role for opioid receptors in the regulation of Pavlovian association formation. , 2004, Behavioral neuroscience.

[12]  M. Kalish,et al.  Connectionism: A Hands-On Approach, Michael R.W. Dawson. Blackwell (2005), £50.00 (hbk)/£19.99 (pbk), (200 pp.), ISBN: 1 405 13074 1 (hbk)/1 405 12807 0 , 2006 .

[13]  Philipp Slusallek,et al.  Introduction to real-time ray tracing , 2005, SIGGRAPH Courses.

[14]  R. Rescorla.  Reduction in the effectiveness of reinforcement after prior excitatory conditioning , 1970 .

[15]  A G Barto,et al.  Toward a modern theory of adaptive networks: expectation and prediction. , 1981, Psychological review.

[16]  J. Pearce.  Animal Learning and Cognition: An Introduction , 1997 .

[17]  J. Pearce.  Evaluation and development of a connectionist theory of configural learning , 2002, Animal learning & behavior.

[18]  Michael R. W. Dawson,et al.  Modifying the Generalized Delta Rule to Train Networks of Non-monotonic Processors for Pattern Classification , 1992 .

[19]  N. Mackintosh,et al.  An elemental model of associative learning: I. Latent inhibition and perceptual learning , 2000 .

[20]  J. Kruschke.  Toward a unified model of attention in associative learning , 2001 .

[21]  L. Kamin.  Attention-like processes in classical conditioning , 1967 .

[22]  D. Danks.  Equilibria of the Rescorla--Wagner model , 2003 .

[23]  N. Mackintosh,et al.  Associative learning and elemental representation: II. Generalization and discrimination , 2002, Animal learning & behavior.

[24]  Marcia L. Spetch,et al.  The implications of null patterns and output unit activation functions on simulation studies of learning: A case study of patterning , 2005 .

[25]  R. Rescorla,et al.  A theory of Pavlovian conditioning : Variations in the effectiveness of reinforcement and nonreinforcement , 1972 .

[26]  R. Rescorla.  Summation and overexpectation with qualitatively different outcomes , 1999 .

[27]  Norbert Jankowski,et al.  Survey of Neural Transfer Functions , 1999 .