Arrow Research search
Back to NMR

NMR 2025

Toward Defeasible Reasoning Using Knowledge Compilation Techniques

Conference Paper Full Papers Artificial Intelligence · Knowledge Representation · Logic in Computer Science

Abstract

We explore the use of ordered binary decision diagrams (OBDDs) and multi-terminal binary decision diagrams (MTBDDs) for compiling and executing defeasible reasoning, focusing on System Z, also known as rational closure. We introduce an OBDD-based algorithm for System Z inference that parallels traditional SAT-based approaches. Additionally, we investigate the use of MTBDDs for encoding ordinal conditional functions (OCFs), providing a more general and unified knowledge compilation framework for defeasible reasoning tasks involving ranking functions. While our primary focus is on System Z, the proposed techniques are applicable to other forms of defeasible reasoning and conditional logic more broadly. Experimental results demonstrate that OBDDs can significantly accelerate inference. However, our evaluation is limited to synthetically generated belief bases, and further research is needed to validate benefits in real-world scenarios. Overall, our findings suggest that binary decision diagrams offer a promising alternative to SAT-based methods for defeasible reasoning, and warrant further investigation.

Authors

Keywords

No keywords are indexed for this paper.

Context

Venue
International Workshop on Non-Monotonic Reasoning
Archive span
1984-2025
Indexed papers
227
Paper id
1015405412482820184
v2026.09.13