Plan-then-compile architectures
暂无分享,去创建一个
Several architectures have recently been developed which integrate planning, reacting, and knowledge compilation learning, in order to improve system reactivity while retaining system flexibility. This paper assesses strengths and weaknesses of such Plan-Then-Compile architectures, by analyzing one such architecture, Theo-Agent [Mitchell 90]. We describe capabilities and limitations of the Theo-Agent, and use these to try to understand the generic capabilities and limits of Plan-Then-Compile architectures more generally.
[1] Nils J. Nilsson,et al. Artificial Intelligence , 1974, IFIP Congress.
[2] John E. Laird,et al. Integrating, Execution, Planning, and Learning in Soar for External Environments , 1990, AAAI.
[3] Kurt VanLehn,et al. The Content of Self-Explanations , 1990 .
[4] Tom M. Mitchell,et al. Becoming Increasingly Reactive , 1990, AAAI.
[5] Oren Etzioni,et al. Explanation-Based Learning: A Problem Solving Perspective , 1989, Artif. Intell..