Qualitative Reasoning for Intelligent Agents
Sponsor: Computer Science Division, Office of Naval Research
Principal Investigator: Kenneth D. Forbus
Project Summary: This project explores the use of qualitative physics to provide capabilities for intelligent agents. Understanding and using common sense reasoning about the physical world is a necessary prerequisite to creating many kinds of useful intelligent agents that collaborate with human partners to accomplish tasks. Examples of such tasks include damage control assessment, operations planning, sifting through on-line information for relevant data, teaching and tutoring, and developing complex scientific and engineering models. Towards these ends, we are
- Exploring methods for integrating first-principles and case-based qualitative reasoning, to achieve anytime-performance and to develop agents that can incrementally improve their expertise, based on accumulated experience.
- Developing qualitative representations and reasoning techniques for natural language semantics, to support the creation of broad-based language systems for agents that must communicate about physical situations and systems. Examples include understanding explanations of how systems work and damage control reports.
- Developing techniques for performing back of the envelope estimations and simulations, to support agents that must make quick quantitative estimates to help solve problems under time pressure with partial and conflicting information. (Examples include estimating how long it will take to pump out a flooded compartment and estimating how much oxygen is left on MIR).
Our vehicle for these investigations is the creation of an experimental prototype, an Explanation Agent that accumulates explanations of how engineered systems work, and that uses this accumulated knowledge to answer questions and interactively formulate task-specific models of those systems.
Selected publications:
Paritosh, P.K. and Forbus, K.D. (2001). Common sense on the envelope. Proceedings of the 15th International Workshop on Qualitative Reasoning.
Kuehne, S.E. and Forbus, K. D. (2002). Qualitative physics as a component in natural language semantics: A progress report. Proceedings of the Twenty-fourth Annual Meeting of the Cognitive Science Society, George Mason University, Fairfax, VA.
Paritosh, P.K. (2003). A sketch of a theory of quantity. Proceedings of the 17th International Workshop on Qualitative Reasoning, Brasilia, Brazil, August.
Kuehne, S. and Forbus, K. (2004). Capturing QP-relevant information from natural language text. Proceedings of the 18th International Qualitative Reasoning Workshop, Evanston, Illinois, USA, August.
Paritosh, P. and Forbus, K. (2004). Using strategies and AND/OR decomposition for back of the envelope reasoning. Proceedings of the 18th International Qualitative Reasoning Workshop, Evanston, Illinois, August.
Kuehne, S. E. (2004). On the Representation of Physical Quantities in Natural Language Text. Proceedings of the Twenty-sixth Annual Meeting of the Cognitive Science Society, Chicago, Illinois, August.
Paritosh, P. and Forbus, K. (2005). Analysis of Strategic Knowledge in Back of the Envelope Reasoning. In Proceedings of the 20th National Conference on Artificial Intelligence (AAAI-05), Pittsburgh, PA.