UT ML Group: Theory and Knowledge Refinement

Most machine learning does not exploit prior knowledge. Theory refinement (a.k.a. theory revision or knowledge-base refinement) is the task of modifying an initial imperfect knowledge-base (KB) to make it consistent with empirical data. The goal is to improve the performance of learning by exploiting prior knowledge, and to acquire knowledge which is more comprehensible and related to existing concepts in the domain. Another motivation is to automate the process of knowledge refinement in the development of expert systems and other knowledge-based systems.

We have developed systems for refining knowledge in various forms including:

These systems have demonstrated an ability to revise real knowledge bases and improve learning in several real-world domains including:

Publications

  1. Guiding a Reinforcement Learner with Natural Language Advice: Initial Results in RoboCup Soccer [Abstract] [PDF]
    Gregory Kuhlmann, Peter Stone, Raymond J. Mooney, and Jude W. Shavlik
    Proceedings of the AAAI-2004 Workshop on Supervisory Control of Learning and Adaptive Systems, pp. 30-35, San Jose, CA, July 2004.

  2. Theory Refinement for Bayesian Networks with Hidden Variables [Abstract] [PDF]
    Sowmya Ramachandran and Raymond J. Mooney
    Proceedings of the Fifteenth International Conference on Machine Learning (ICML-98), Madison, WI, pp. 454-462, July 1998.

  3. Theory Refinement of Bayesian Networks with Hidden Variables [Abstract] [PDF]
    Sowmya Ramachandran and Raymond J. Mooney
    Ph.D. Thesis, Department of Computer Sciences, University of Texas at Austin, May 1998.
    139 pages
    Also appears as Technical Report AI 98-265, Artificial Intelligence Lab, University of Texas at Austin.

  4. Integrating Abduction and Induction in Machine Learning [Abstract] [PDF]
    Raymond J. Mooney
    Working Notes of the IJCAI-97 Workshop on Abduction and Induction in AI, Nagoya, Japan, pp. 37-42, August 1997.

  5. Parameter Revision Techniques for Bayesian Networks with Hidden Variables: An Experimental Comparison [Abstract] [PDF]
    Sowmya Ramachandran and Raymond J. Mooney
    Unpublished Technical Note, January 1997.

  6. A Novel Application of Theory Refinement to Student Modeling [Abstract] [PDF]
    Paul Baffes and Raymond J. Mooney
    Best Paper Award
    Proceedings of the Thirteenth National Conference on Artificial Intelligence (AAAI-96), pp. 403-408, Portland, OR, August 1996.

  7. Combining Symbolic and Connectionist Learning Methods to Refine Certainty-Factor Rule-Bases [Abstract] [PDF]
    J. Jeffrey Mahoney
    Ph.D. Thesis, Department of Computer Sciences, University of Texas at Austin, May 1996.
    113 pages

  8. Refinement-Based Student Modeling and Automated Bug Library Construction [Abstract] [PDF]
    Paul Baffes and Raymond Mooney
    Journal of Artificial Intelligence in Education, 7, 1 (1996), pp. 75-116.

  9. Revising Bayesian Network Parameters Using Backpropagation [Abstract] [PDF]
    Sowmya Ramachandran and Raymond J. Mooney
    Proceedings of the International Conference on Neural Networks (ICNN-96), Special Session on Knowledge-Based Artificial Neural Networks, pp. 82-87, Washington DC, June 1996.

  10. Refinement of Bayesian Networks by Combining Connectionist and Symbolic Techniques [Abstract] [PDF]
    Sowmya Ramachandran
    Ph.D. proposal, Department of Computer Sciences, University of Texas at Austin, October 1995.
    34 pages

  11. Automated Refinement of First-Order Horn-Clause Domain Theories [Abstract] [PDF]
    Bradley L. Richards and Raymond J. Mooney
    Machine Learning 19,2 (1995), pp. 95-131.

  12. A Preliminary PAC Analysis of Theory Revision [Abstract] [PDF]
    Raymond J. Mooney
    Computational Learning Theory and Natural Learning Systems, Vol. 3, T. Petsche, S. Judd, and S. Hanson (Eds.), MIT Press, 1995, pp. 43-53.

  13. Automatic Student Modeling and Bug Library Construction using Theory Refinement [Abstract] [PDF]
    Paul T. Baffes
    Ph.D. Thesis, Department of Computer Sciences, University of Texas at Austin, December 1994.
    211 pages
    Also appears as Technical Report AI 94-215, Artificial Intelligence Laboratory, University of Texas at Austin, February 1994.

  14. Comparing Methods For Refining Certainty Factor Rule-Bases [Abstract] [PDF]
    J. Jeffrey Mahoney and Raymond J. Mooney
    Proceedings of the Eleventh International Workshop on Machine Learning (ML-94), pp. 173-180, Rutgers, NJ, July 1994.

  15. Modifying Network Architectures For Certainty-Factor Rule-Base Revision [Abstract] [PDF]
    J. Jeffrey Mahoney and Raymond J. Mooney
    Proceedings of the International Symposium on Integrating Knowledge and Neural Heuristics 1994 (ISIKNH-94), pp. 75-85, Pensacola, FL, May 1994.

  16. A Multistrategy Approach to Theory Refinement [Abstract] [PDF]
    Raymond J. Mooney and Dirk Ourston
    Machine Learning: A Multistrategy Approach, Vol. IV, R.S. Michalski & G. Teccuci (Eds.), pp.141-164, Morgan Kaufman, San Mateo, CA, 1994.

  17. Theory Refinement Combining Analytical and Empirical Methods [Abstract] [PDF]
    Dirk Ourston and Raymond J. Mooney
    Artificial Intelligence, 66 (1994), pp. 311--344.

  18. Extending Theory Refinement to M-of-N Rules [Abstract] [PDF]
    Paul T. Baffes and Raymond J. Mooney
    Informatica, 17 (1993), pp. 387-397.

  19. Symbolic Revision of Theories With M-of-N Rules [Abstract] [PDF]
    Paul T. Baffes and Raymond J. Mooney
    Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence (IJCAI-93), pp. 1135-1140, Chambéry, France, July 1993.

  20. Learning to Model Students: Using Theory Refinement to Detect Misconceptions [Abstract] [PDF]
    Paul T. Baffes
    Ph.D. proposal, Department of Computer Sciences, University of Texas at Austin, June 1993.
    34 pages

  21. Combining Connectionist and Symbolic Learning to Refine Certainty-Factor Rule-Bases [Abstract] [PDF]
    J. Jeffrey Mahoney and Raymond J. Mooney
    Connection Science, 5 (1993), pp. 339-364. (Special issue on Architectures for Integrating Neural and Symbolic Processing)

  22. Integrating Theory and Data in Category Learning [Abstract] [PDF]
    Raymond J. Mooney
    Categorization by Humans and Machines: The Psychology of Learning and Motivation, Vol. 29, G. Nakamura, R. Taraban, & D.L. Medin (Eds.), pp. 189-218, Academic Press, Orlando, FL, 1993.

  23. Induction Over the Unexplained: Using Overly-General Domain Theories to Aid Concept Learning [Abstract] [PDF]
    Raymond J. Mooney
    Machine Learning, 10, 1 (1993), pp. 79-110.

  24. Using Theory Revision to Model Students and Acquire Stereotypical Errors [Abstract] [PDF]
    Paul T. Baffes and Raymond J. Mooney
    Proceedings of the Fourteenth Annual Conference of the Cognitive Science Society, pp. 617-622, Bloomington, IN, July 1992.

  25. Combining Symbolic and Neural Learning to Revise Probabilistic Theories [Abstract] [PDF]
    J. Jeffrey Mahoney and Raymond J. Mooney
    Proceedings of the 1992 Machine Learning Workshop on Integrated Learning in Real Domains, Aberdeen, Scotland, July 1992.

  26. Automated Debugging of Logic Programs via Theory Revision [Abstract] [PDF]
    Raymond J. Mooney and Bradley L. Richards
    Proceedings of the Second International Workshop on Inductive Logic Programming, Tokyo, Japan, June 1992.

  27. Batch versus Incremental Theory Refinement [Abstract] [PDF]
    Raymond J. Mooney
    Proceedings of the 1992 AAAI Spring Symposium on Knowledge Assimilation, Standford, CA, March 1992.

  28. Improving Shared Rules in Multiple Category Domain Theories [Abstract] [PDF]
    Dirk Ourston and Raymond J. Mooney
    Proceedings of the Eighth International Machine Learning Workshop, pp. 534-538, Evanston, IL, June 1991.

  29. Constructive Induction in Theory Refinement [Abstract] [PDF]
    Raymond J. Mooney and Dirk Ourston
    Proceedings of the Eighth International Machine Learning Workshop, pp. 178-182, Evanston, IL, June 1991.

  30. Theory Refinement with Noisy Data [Abstract] [PDF]
    Raymond J. Mooney and Dirk Ourston
    Technical Report AI 91-153, Artificial Intelligence Lab, University of Texas at Austin, March 1991.

  31. Changing the Rules: A Comprehensive Approach to Theory Refinement [Abstract] [PDF]
    Ourston, D. and Mooney, R. J.
    Proceedings of the Eighth National Conference on Artificial Intelligence, pp. 815-820, Boston, MA, August 1990.

mooney@cs.utexas.edu