KDMAS: A Multi-Agent System for Knowledge Discovery via Planning

In the real world, there are some domain knowledge discovery problems that can be formulated into knowledge-based planning problems, such as chemical reaction process and biological pathway discovery problems (Khan et al. 2003). A view of these domain problems can be re-cast as a planning problem, such that initial and final states are known and processes can be captured as abstract operators that modify the environment. For example, approaching biological pathway discovery with an AI planning approach would mean that a valid plan that transfers the initial state into the goal state is a hypothetical pathway that prescribes the order of events that must occur to effect the goal state. We believe that AI planning technology can provide a modeling formalism for this task such that hypotheses can be generated, tested, queried and qualitatively simulated to improve the domain knowledge and rules. Our current approach is to build toward a general multiagent system for knowledge discovery (KDMAS) via planning. The plans produced are hypotheses capturing relevant qualitative information regarding domain knowledge. We will use the biological pathway domain as an application to present our approach, focusing on acquiring planning knowledge and hypothesis generation. As shown in Figure 1, KDMAS integrates: The domain knowledge, which is built up with the domain experts’ aid, and can be translated into the planning formalism automatically by the planning formalism agent. Other domain knowledge from data that is retrieved from heterogeneous sources including human experts by information gathering multi agents and used for hypothesis generation. Computer interpretable plans capturing relevant qualitative information regarding the domain problems such as signal transduction pathways. Testable hypotheses regarding gaps in knowledge of the domain such as a biological pathway to drive future research.