Learning functional descriptions from examples Public Deposited

http://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/rj430687s

Descriptions

Attribute NameValues
Creator
Abstract or Summary
  • The task of inductive learning from examples places constraints on the representation of training instances and concepts. These constraints are different from, and often incompatible with, the constraints placed on the representation by the performance task. This incompatibility is severe when learning functional concepts and explains why previous researchers have found it so difficult to construct good representations for inductive learningthey were trying to achieve a compromise between these two sets of constraints. This thesis addresses this problem, and takes a different approach. Rather than designing a compromise representation we employ two different representations: one for learning and one for performance. The system developed learns concepts in chess and checkers. Training instances are presented in the "performance representation" as simple board positions, then converted to the "learning representation" via a search process that builds an explanation of the outcome of the position. Inductive generalization is performed over these explanations to form descriptions of the concepts in terms of the moves and goals involved. Finally the concepts are translated back into the "performance representation" to support efficient recognition of future instances. The advantages of this "two representation" approach are (a) many fewer training instances are required to learn the concept, (b) the biases of the learning program are very simple, and (c) the learning system requires virtually no "vocabulary engineering" to learn concepts in a new domain.
Resource Type
Date Available
Date Copyright
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Academic Affiliation
Non-Academic Affiliation
Subject
Rights Statement
Peer Reviewed
Language
Digitization Specifications
  • File scanned at 300 ppi (Monochrome) using ScandAll PRO 1.8.1 on a Fi-6770A in PDF format. CVista PdfCompressor 5.0 was used for pdf compression and textual OCR.
Replaces
Additional Information
  • description.provenance : Approved for entry into archive by Patricia Black(patricia.black@oregonstate.edu) on 2013-07-10T17:18:17Z (GMT) No. of bitstreams: 1 FlannNicholasS1986.pdf: 1344162 bytes, checksum: c4083e7d6579402fa3dbafe08eb70f47 (MD5)
  • description.provenance : Submitted by Sergio Trujillo (jstscanner@gmail.com) on 2013-07-09T21:36:05Z No. of bitstreams: 1 FlannNicholasS1986.pdf: 1344162 bytes, checksum: c4083e7d6579402fa3dbafe08eb70f47 (MD5)
  • description.provenance : Approved for entry into archive by Patricia Black(patricia.black@oregonstate.edu) on 2013-07-15T14:32:47Z (GMT) No. of bitstreams: 1 FlannNicholasS1986.pdf: 1344162 bytes, checksum: c4083e7d6579402fa3dbafe08eb70f47 (MD5)
  • description.provenance : Made available in DSpace on 2013-07-15T14:32:47Z (GMT). No. of bitstreams: 1 FlannNicholasS1986.pdf: 1344162 bytes, checksum: c4083e7d6579402fa3dbafe08eb70f47 (MD5) Previous issue date: 1986-05-01

Relationships

Parents:

This work has no parents.

Last modified

Downloadable Content

Download PDF

Items