The Anatomy of a General Learning Mechanism

In this article we describe an approach to the construction of a general learning mechanism based on chunking in Soar. Chunking is a learning mechanism that acquires rules from goal-based experience. Soar is a general problem-solving architecture with a rulebased memory. In previous work we have demonstrated how the combination of chunking and Soar could acquire search-control knowledge (strategy acquisition) and operator implementation rules in both search-based puzzle tasks and knowledge-based expert-systems tasks. In this work we examine the anatomy of chunking in Soar and provide a new demonstration of its learning capabilities involving the acquisition and use of macro-operators. This paper will appear in Machine Learning, vol. 1 no. 1, January 1986. This paper is also available from the Department of Computer Science, Carnegie-Mellon University, as technical report CMU-CS-85-154 and from the Knowledge Systems Laboratory, Department of Computer Science, Stanford University, as report KSL-85-34. XEROX Xerox Corporation Palo Alto Research Centers 3333 Coyote Hill Road Palo Alto, California 94304 'Departments of Psychology and Computer Science, Stanford University Department of Computer Science, Carnegie-Mellon University