论文信息 - The DUCTOR: A Theory Revision System for Propositional Domains

The DUCTOR: A Theory Revision System for Propositional Domains

This paper presents the theory revision component of an integrated learning system called the DUCTOR. This component is capable of repairing imperfect domain theories by generalizing and/or specializing the rules responsible for incorrect classifications of examples. These repairs are made so as to minimize the changes to the original domain theory that will make the theory consistent with the set of examples. The theory revision system has been tested on a real-world medical domain theory, with very good results.

Timothy Cain | Timothy Cain

[1] Michael R. Lowry,et al. Learning Physical Descriptions From Functional Definitions, Examples, and Precedents , 1983, AAAI.

[2] Geoffrey G. Towell,et al. Symbolic knowledge and neural networks: insertion, refinement and extraction , 1992 .

[3] Allen Ginsberg,et al. Theory Reduction, Theory Revision, and Retranslation , 1990, AAAI.

[4] Raymond J. Mooney,et al. Changing the Rules: A Comprehensive Approach to Theory Refinement , 1990, AAAI.

[5] Allen Ginsberg,et al. Theory Revision via Prior Operationalization , 1988, AAAI.

[6] Temple F. Smith. Occam's razor , 1980, Nature.

[7] D. Ourston. Using explanation-based and empirical methods in theory revision , 1991 .

[8] Harry E. Pople,et al. Session 6 Theorem Proving and Logic: I I ON THE MECHANIZATION OF ABDUCTIVE LOGIC , 2006 .

[9] Michael J. Pazzani,et al. Integrated Learning with Incorrect and Incomplete Theories , 1988, ML.

[10] Leslie G. Valiant,et al. A theory of the learnable , 1984, STOC '84.

[11] Leslie G. Valiant,et al. A general lower bound on the number of examples needed for learning , 1988, COLT '88.

[12] Allen Ginsberg. Knowledge-Base Reduction: A New Approach to Checking knowledge Bases for Inconsistency and Redundancy , 1988, AAAI.