Case-Based Argumentation via Process Models

May 2003

Proceedings of the Sixteenth International Florida Artificial Intelligence Research Society Conference (FLAIRS 2003)

Volume

Proceedings of the Sixteenth International Florida Artificial Intelligence Research Society Conference (FLAIRS 2003)

All Papers

We introduce AHEAD (Analogical Hypothesis Elaborator for Activity Detection), a software system we are developing for the DARPA EELD (Evidence Extraction and Link Discovery) program. AHEAD performs case-based hypothesis elaboration using process models. We are applying AHEAD, which embodies a domain-independent approach, to elaborate hypothesized hostile activities. In this application, AHEAD is given as input (1) structured evidence and (2) a hypothesis concerning the activities of an asymmetric adversary (e.g., a terrorist organization). The system outputs a detailed symbolic argument supporting and/or opposing the given hypothesis. Combining case- based/analogical reasoning techniques, using the FIRE Analogy Server with qualitative functional processes represented as Task-Method-Knowledge models, AHEAD extracts additional implications of the hypothesis to develop a coherent argument that supports and/or contradicts it. We detail AHEAD’s design, its role in EELD, its implementation status, and our plans for its evaluation.