Klenk, Matthew (Navy Center for Applied Research in Artificial Intelligence) | Aha, David W. (Navy Center for Applied Research in Artificial Intelligence) | Molineaux, Matt (Knexus Research Corporation)
Case-based reasoning (CBR) is a problem-solving process in which a new problem is solved by retrieving a similar situation and reusing its solution. Transfer learning occurs when, after gaining experience from learning how to solve source problems, the same learner exploits this experience to improve performance and/or learning on target problems. In transfer learning, the differences between the source and target problems characterize the transfer distance. CBR can support transfer learning methods in multiple ways. We illustrate how CBR and transfer learning interact and characterize three approaches for using CBR in transfer learning: (1) as a transfer learning method, (2) for problem learning, and (3) to transfer knowledge between sets of problems. We describe examples of these approaches from our own and related work and discuss applicable transfer distances for each. We close with conclusions and directions for future research applying CBR to transfer learning.
This research was motivated by the widely held belief that constructing an automatic program synthesis system that can accept a high-level description of a problem for an arbitrary domain and generate code for the problem completely automatically is pragmatically impossible. However, by focusing on a well-defined domain, it is possible to incorporate sufficient knowledge within a system so that it can communicate with an end user at the level of his(her) application and automatically generate a program from a problem specification. Such knowledge-based systems often employ a catalog of transformational rules that progressively refine an abstract specification into a concrete implementation. A major research issue in such systems is how to increase the efficiency of the systems by controlling the application of rules and avoiding repetitive traversal of the search space. In my Ph.D. dissertation (Bhansali 1991), I develop an integrated knowledge-based framework for efficiently synthesizing programs by bringing together ideas from the fields of software engineering (software reuse, domain modeling) and The knowledge base consists of three subcomponents: a concept dictionary, a library of reusable components, and a layered rule base.
While it is commonly agreed that analogy is useful in human problem solving, exactly how analogy can and should be used remains an intriguing problem. VanLehn (1998) for instance argues that there are differences in how novices and experts use analogy, but the VanLehn and Jones (1993) Cascade model does not implement these differences. This paper analyzes several variations in strategies for using analogy to explore possible sources of novice/expert differences. We describe a series of ablation experiments on an expert model to examine the effects of strategy variations in using analogy in problem solving. We provide evidence that failing to use qualitative reasoning when encoding problems, being careless in validating analogical inferences, and not using multiple retrievals can degrade the efficiency of problem-solving.