Lazy Incremental Learning of Control Knowledge for Efficiently Obtaining Quality Plans

General-purpose generative planners use domain-independent search heuristics to generate solutions for problems in a variety of domains. However, in some situations these heuristics force the planner to perform inefficiently or obtain solutions of poor quality. Learning from experience can help to identify the particular situations for which the domain-independent heuristics need to be overridden. Most of the past learning approaches are fully deductive and eagerly acquire correct control knowledge from a necessarily complete domain theory and a few examples to focus their scope. These learning strategies are hard to generalize in the case of nonlinear planning, where it is difficult to capture correct explanations of the interactions among goals, multiple planning operator choices, and situational data. In this article, we present a lazy learning method that combines a deductive and an inductive strategy to efficiently learn control knowledge incrementally with experience. We present hamlet, a system we developed that learns control knowledge to improve both search efficiency and the quality of the solutions generated by a nonlinear planner, namely prodigy4.0. We have identified three lazy aspects of our approach from which we believe hamlet greatly benefits: lazy explanation of successes, incremental refinement of acquired knowledge, and lazy learning to override only the default behavior of the problem solver. We show empirical results that support the effectiveness of this overall lazy learning approach, in terms of improving the efficiency of the problem solver and the quality of the solutions produced.

[1]  Oren Etzioni,et al.  PRODIGY: an integrated architecture for planning and learning , 1991, SGAR.

[2]  R. Bronson The Knowledge , 2002, Annals of Internal Medicine.

[3]  R. Mooney,et al.  Explanation-Based Learning: An Alternative View , 1986, Machine Learning.

[4]  Tom M. Mitchell,et al.  Explanation-Based Generalization: A Unifying View , 1986, Machine Learning.

[5]  Oren Etzioni,et al.  PRODIGY4.0: The Manual and Tutorial , 1992 .

[6]  Ray Bareiss,et al.  Concept Learning and Heuristic Classification in WeakTtheory Domains , 1990, Artif. Intell..

[7]  Manuela Veloso Nonlinear problem solving using intelligent casual-commitment , 1989 .

[8]  S. Kambhampati,et al.  Learning Explanation-Based Search Control Rules for Partial Order Planning , 1994, AAAI.

[9]  Pat Langley,et al.  Learning Effective Search Heuristics , 1983, IJCAI.

[10]  Allen Newell,et al.  Chunking in Soar , 1986 .

[11]  Raymond J. Mooney,et al.  Combining FOIL and EBG to Speed-up Logic Programs , 1993, IJCAI.

[12]  Oren Etzioni,et al.  DYNAMIC: A New Role for Training Problems in EBL , 1992, ML.

[13]  D. Kibler,et al.  Instance-based learning algorithms , 2004, Machine Learning.

[14]  N. Nilsson,et al.  Readings in Artificial Intelligence , 1981 .

[15]  James A. Hendler,et al.  Massively parallel support for case-based planning , 1993, Proceedings of 9th IEEE Conference on Artificial Intelligence for Applications.

[16]  William W. Cohen Learning Approximate Control Rules of High Utility , 1990, ML.

[17]  Lawrence Birnbaum,et al.  Proceedings of the eighth international workshop on Machine learning , 1991 .

[18]  Leslie G. Valiant,et al.  A theory of the learnable , 1984, CACM.

[19]  J. R. Quinlan Learning Logical Definitions from Relations , 1990 .

[20]  Nils J. Nilsson,et al.  Artificial Intelligence , 1974, IFIP Congress.

[21]  Neeraj Bhatanagar Learning by incomplete explanations of failures in recursive domains , 1992, ICML 1992.

[22]  Manuela M. Veloso,et al.  The Need for Different Domain-independent Heuristics , 1994, AIPS.

[23]  John R. Anderson,et al.  Machine learning - an artificial intelligence approach , 1982, Symbolic computation.

[24]  Ingrid Zukerman,et al.  Learning Search Control Rules for Planning: An Inductive Approach , 1991, ML.

[25]  Manuela M. Veloso,et al.  Flexible Strategy Learning: Analogical Replay of Problem Solving Episodes , 1994, AAAI.

[26]  Manuela M. Veloso,et al.  Learning strategy knowledge incrementally , 1994, Proceedings Sixth International Conference on Tools with Artificial Intelligence. TAI 94.

[27]  David Ruby,et al.  Learning Episodes for Optimization , 1992, ML.

[28]  Jaime G. Carbonell,et al.  Learning effective search control knowledge: an explanation-based approach , 1988 .

[29]  Manuela M. Veloso,et al.  Linkability: Examining Causal Link Commitments in Partial-order Planning , 1994, AIPS.

[30]  Prasad Tadepalli,et al.  Lazy ExplanationBased Learning: A Solution to the Intractable Theory Problem , 1989, IJCAI.

[31]  Jaime G. Carbonell,et al.  Control Knowledge to Improve Plan Quality , 1994, AIPS.

[32]  Richard Fikes,et al.  Learning and Executing Generalized Robot Plans , 1993, Artif. Intell..

[33]  Alan Bundy,et al.  Proceedings of the Eighth International Joint Conference on Artificial Intelligence , 1983 .

[34]  Kristian J. Hammond,et al.  Case-Based Planning: Viewing Planning as a Memory Task , 1989 .

[35]  Manuela M. Veloso,et al.  Incremental Learning of Control Knowledge for Nonlinear Problem Solving , 1994, ECML.

[36]  Subbarao Kambhampati,et al.  Explanation-Based Generalization of Partially Ordered Plans , 1991, AAAI.

[37]  Allen Newell,et al.  Chunking in Soar: The anatomy of a general learning mechanism , 1985, Machine Learning.

[38]  Pavel Brazdil,et al.  Proceedings of the European Conference on Machine Learning , 1993 .

[39]  James A. Hendler,et al.  Massively parallel support for case-based planning , 1994, IEEE Expert.

[40]  STEVEN MINTON,et al.  A reply to Zito-Wolf's book review ofLearning search control knowledge: An explanation-based approach , 2004, Machine Learning.

[41]  Tara A. Estlin,et al.  Hybrid learning of search control for partial-order planning , 1996 .

[42]  Tom M. Mitchell,et al.  Learning by experimentation: acquiring and refining problem-solving heuristics , 1993 .

[43]  Oren Etzioni,et al.  Why EBL Produces Overly-Specific Knowledge: A Critique of the PRODIGY Approaches , 1992, ML.

[44]  David L. Waltz,et al.  Toward memory-based reasoning , 1986, CACM.

[45]  Eugene Fink,et al.  Integrating planning and learning: the PRODIGY architecture , 1995, J. Exp. Theor. Artif. Intell..

[46]  Daniel S. Weld,et al.  A Domain-Independent Algorithm for Plan Adaptation , 1994, J. Artif. Intell. Res..

[47]  J. Ross Quinlan,et al.  Learning logical definitions from relations , 1990, Machine Learning.

[48]  James A. Hendler,et al.  Flexible reuse and modification in hierarchical planning: a validation structure-based approach , 1989 .

[49]  Richard Waldinger,et al.  Achieving several goals simultaneously , 1977 .

[50]  Peter Clark,et al.  Lazy Partial Evaluation: An Integration of Explanation-Based Generalization and Partial Evaluation , 1992, ML.

[51]  Manuela M. Veloso,et al.  Planning and Learning by Analogical Reasoning , 1994, Lecture Notes in Computer Science.