Machine Learning for General Problem Solving

摘要

Part I describes progress in research relevant to problem solving and learning using trigonometric identities and logical equations. The work previously described (AD-608 544) dealing with trigonometric proof learning has been analyzed and extended with the general solution and learning approach applied to proofs of logical equations. Also included is a brief description of work which investigates concept formation modeled in one sense after the work of Piaget. Part II relates the research in tree pruning as related to computers and game playing. Recent work has emphasized learning processes independent of scoring functions. Comparison is made between human play and machine play in specific game environments. Part II of the report describes evaluation of the research goal of extending the problem solving and learning processes into the field of integral calculus. It was determined, after considerable experimentation and research, that more useful research expenditure could be made in investigations of other learning areas. (Author)