Generating Decision Structures and Causal Explanations for Decision Making
This addresses the challenge of enabling autonomous agents like robots to make decisions and explain outcomes, but it is incremental as it builds on existing causal reasoning methods.
The paper tackles the problem of generating decision structures and causal explanations from unstructured knowledge databases, showing that these tasks are intractable without constraints and proposing causal theories to improve efficiency, with a program demonstrating increased efficiency when causal constraints are applied.
This paper examines two related problems that are central to developing an autonomous decision-making agent, such as a robot. Both problems require generating structured representafions from a database of unstructured declarative knowledge that includes many facts and rules that are irrelevant in the problem context. The first problem is how to generate a well structured decision problem from such a database. The second problem is how to generate, from the same database, a well-structured explanation of why some possible world occurred. In this paper it is shown that the problem of generating the appropriate decision structure or explanation is intractable without introducing further constraints on the knowledge in the database. The paper proposes that the problem search space can be constrained by adding knowledge to the database about causal relafions between events. In order to determine the causal knowledge that would be most useful, causal theories for deterministic and indeterministic universes are proposed. A program that uses some of these causal constraints has been used to generate explanations about faulty plans. The program shows the expected increase in efficiency as the causal constraints are introduced.