Instantiation-based Formalization of Logical Reasoning Tasks using Language Models and Logical Solvers
This work addresses the problem of unreliable AI-driven reasoning systems for applications requiring dependable and autonomous reasoning, though it appears incremental as it builds on existing methods with logical solvers.
The paper tackles the challenge of improving reasoning robustness in large language models by introducing Semantic Self-Verification (SSV), a method that combines language models with logical solvers to formalize reasoning tasks, resulting in significantly advanced reasoning accuracy and near-perfect precision in verification over open benchmarks.
Robustness of reasoning remains a significant challenge for large language models, and addressing it is essential for the practical applicability of AI-driven reasoning systems. We introduce Semantic Self-Verification (SSV), a novel approach that addresses the key challenge in combining language models with the rigor of logical solvers: to accurately formulate the reasoning problem from natural language to the formal language of the solver. SSV uses a consistency-based approach to produce strong abstract formalizations of problems using concrete instantiations that are generated by the model and verified by the solver. In addition to significantly advancing the overall reasoning accuracy over the state-of-the-art, a key novelty that this approach presents is a feature of verification that has near-perfect precision over a significant coverage of cases, as we demonstrate on open reasoning benchmarks. We propose such *near-certain reasoning* as a new approach to reduce the need for manual verification in many cases, taking us closer to more dependable and autonomous AI reasoning systems.