Running head : Long-Term Symbolic Learning Long-Term Symbolic Learning

What are the characteristics of long-term learning? We investigated the characteristics of long-term, symbolic learning using the Soar and ACT-R cognitive architectures running cognitive models of two simple tasks. Long sequences of problems were run collecting data to answer fundamental questions about long-term, symbolic learning. We examined whether symbolic learning continues indefinitely, how the learned knowledge is used, and whether performance degrades over the long term. We report that in both systems, symbolic learning eventually stopped, learned knowledge was used differently in different stages but the resulting production knowledge was used uniformly, and, finally, both Soar and ACT-R do eventually suffer from degraded performance with long-term continuous learning. We also discuss ACT-R implementation and theoretic causes of ACT-R’s performance problems and settings that appear to avoid the performance problems in ACT-R. Kennedy, W.G., & Trafton, J.G. (in press 2007) Long-term Symbolic Learning. Cognitive Systems Research. Available from http://www.mllab.com

[1]  Nils J. Nilsson,et al.  Artificial Intelligence , 1974, IFIP Congress.

[2]  Allen and Rosenbloom Paul S. Newell,et al.  Mechanisms of Skill Acquisition and the Law of Practice , 1993 .

[3]  John R. Anderson.  Acquisition of cognitive skill. , 1982 .

[4]  Patrick Henry Winston,et al.  Artificial intelligence (2nd ed.) , 1984 .

[5]  G. Logan.  Toward an instance theory of automatization. , 1988 .

[6]  Shaul Markovitch,et al.  The Role of Forgetting in Learning , 1988, ML.

[7]  Raymond J. Mooney,et al.  The Effect of Rule Use on the Utility of Explanation-Based Learning , 1989, IJCAI.

[8]  John R. Anderson.  The Adaptive Character of Thought , 1990 .

[9]  Lawrence B. Holder,et al.  The General Utility Problem in Machine Learning , 1990, ML.

[10]  Steven Minton,et al.  Quantitative Results Concerning the Utility of Explanation-based Learning , 1988, Artif. Intell..

[11]  Richard Reviewer-Granger.  Unified Theories of Cognition , 1991, Journal of Cognitive Neuroscience.

[12]  Henrik Boström.  Eliminating redundancy in explanation-based learning , 1992, ICML 1992.

[13]  Allen Newell,et al.  The chunking of goal hierarchies: a generalized model of practice , 1993 .

[14]  John R. Anderson,et al.  Learning and Memory: An Integrated Approach , 1994 .

[15]  John E. Laird,et al.  Coordinated Behavior of Computer Generated Forces in TacAir-Soar , 1994 .

[16]  David C. Wilson,et al.  Learning to Improve Case Adaption by Introspective Reasoning and CBR , 1995, ICCBR.

[17]  Robert B. Doorenbos.  Production Matching for Large Learning Systems , 1995 .

[18]  C. Lebiere,et al.  The Atomic Components of Thought , 1998 .

[19]  C. Lebiere,et al.  An integrated theory of list memory. , 1998 .

[20]  Christian Lebiere,et al.  The dynamics of cognition: An ACT-R model of cognitive arithmetic , 1999, Kognitionswissenschaft.

[21]  John R. Anderson.  Learning and memory: An integrated approach, 2nd ed. , 2000 .

[22]  J. Gregory Trafton,et al.  Memory for goals: an activation-based model , 2002, Cogn. Sci..

[23]  Niels Taatgen,et al.  A model of individual differences in skill acquisition in the Kanfer–Ackerman air traffic control task , 2002, Cognitive Systems Research.

[24]  Kenneth A. De Jong,et al.  Characteristics of Long-term Learning in Soar and its Application to the Utility Problem , 2003, ICML.

[25]  Frank J. Lee,et al.  Production Compilation: A Simple Mechanism to Model Complex Skill Acquisition , 2003, Hum. Factors.

[26]  Glenn A. Iba,et al.  A Heuristic Approach to the Discovery of Macro-Operators , 1989, Machine Learning.

[27]  John R Anderson,et al.  An integrated theory of the mind. , 2004, Psychological review.

[28]  Allen Newell,et al.  The problem of expensive chunks and its solution by restricting expressiveness , 1993, Machine Learning.

[29]  Niels Taatgen,et al.  Modeling Parallelization and Flexibility Improvements in Skill Acquisition: From Dual Tasks to Complex Dynamic Skills , 2005, Cogn. Sci..

[30]  John R. Anderson,et al.  From recurrent choice to skill learning: a reinforcement-learning model. , 2006, Journal of experimental psychology. General.

[31]  Michael D. Byrne,et al.  Modeling the Visual Search of Displays: A Revised ACT-R Model of Icon Search Based on Eye-Tracking Data , 2006, Hum. Comput. Interact..

[32]  G. Gunzelmann.  Understanding Similarities in Performance on Different Orientation Tasks: Strategy Adaptation , 2006 .

[33]  Dario D. Salvucci.  Modeling Driver Behavior in a Cognitive Architecture , 2006, Hum. Factors.