Method for Utilization of Previous Experience in Design Expert Systems

Publisher Summary This chapter uses analogical reasoning in concept learning for utilization of previous experience in design expert systems. The main issue in this approach is: how to automatically recognize an analogy between a source and a target, and how to apply it to generating hypotheses for the target domain. Analogical reasoning is a crucial research area in artificial intelligence as a technique to reason from incomplete knowledge. It promises to overcome the explosive search complexity of finding solutions to novel problems or inducing generalized knowledge from experiences. This chapter presents a technique that uses taxonomic information in the knowledge base for this purpose. This chapter proposes a machine learning technique called inductive prediction by analogy (IPA) that learns the concept-description of a target concept using an analogy between the target and a source similar to the target. A basic idea of IPA is that it uses concept similarities to select an appropriate analogous source using taxonomic information represented by first-order predicate logic. Taxonomic information describes classification of symbols of a knowledge representation language. In a logical framework, the symbols are constants to represent predicates, functions, and constant terms in sentences. IPA is based on an analogy as a mapping between constant symbols of a source predicate and a target predicate. This chapter presents the framework of IPA and defines analogy using taxonomic information. It describes IPA in detail and gives an algorithm of IPA. It also discusses IPA in logic programming and a classification problem in molecular biology. This chapter closes with a discussion on the usefulness of IPA and related work.

[1]  Rogers P. Hall,et al.  Computational Approaches to Analogical Reasoning: A Comparative Analysis , 1989, Artif. Intell..

[2]  David Page,et al.  Generalization with Taxonomic Information , 1990, AAAI.

[3]  Stuart J. Russell,et al.  A Logical Approach to Reasoning by Analogy , 1987, IJCAI.

[4]  Michael R. Genesereth,et al.  Logical foundations of artificial intelligence , 1987 .

[5]  T. Ishikawa.  How to predict it : Inductive Prediction by Analogy Using Taxonomic Information , 1996 .

[6]  Takashi Ishikawa,et al.  Analogy by abstraction: Case retrieval and adaptation for inventive design expert systems , 1996 .

[7]  D. Gentner.  Structure‐Mapping: A Theoretical Framework for Analogy* , 1983 .

[8]  Russell Greiner,et al.  Learning by Understanding Analogies , 1986, Artif. Intell..

[9]  Stephen Muggleton,et al.  Machine Invention of First Order Predicates by Inverting Resolution , 1988, ML.

[10]  R. Henderson,et al.  Model for the structure of bacteriorhodopsin based on high-resolution electron cryo-microscopy. , 1990, Journal of molecular biology.

[11]  A. Bairoch,et al.  The SWISS-PROT protein sequence data bank: current status. , 1994, Nucleic acids research.

[12]  Makoto Haraguchi,et al.  Analogical Logic Program Synthesis from Examples , 1995, ECML.

[13]  Georg E. Schulz,et al.  Principles of Protein Structure , 1979 .

[14]  Siegfried Bell,et al.  Analogical Reasoning for Logic Programming , 1991, EWSL.

[15]  Gheorghe Tecuci,et al.  Plausible justification trees: A framework for deep and dynamic integration of learning strategies , 1993, Machine Learning.

[16]  Maurice Bruynooghe,et al.  Interactive Concept-Learning and Constructive Induction by Analogy , 1992, Machine Learning.