FLARE: Faithful Logic-Aided Reasoning and Exploration

2024-10-14

arXiv.org

AI Summary

Existing large language models (LLMs) frequently exhibit inconsistency between final answers and intermediate reasoning chains in complex reasoning tasks: chain-of-thought (CoT) reasoning is prone to errors, while faithful CoT improves accuracy but relies on task-specific fine-tuning and suffers from poor generalization. This paper introduces the first verifiable reasoning framework that requires no external symbolic solver. It decomposes tasks to guide LLM-based planning, softly formalizes reasoning steps as logic programs, and performs multi-hop exhaustive search within a well-defined semantic space. The framework further enables quantitative assessment of reasoning faithfulness and precise localization of critical decision paths. Its core integrates LLM-driven planning, soft logic-programming formalization, verifiable search, and faithfulness evaluation. Evaluated on nine mainstream reasoning benchmarks, it achieves state-of-the-art performance on seven. Empirical analysis confirms a strong positive correlation between reasoning faithfulness and end-task accuracy, and demonstrates accurate identification of optimal reasoning paths.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Modern Question Answering (QA) and Reasoning approaches based on Large Language Models (LLMs) commonly use prompting techniques, such as Chain-of-Thought (CoT), assuming the resulting generation will have a more granular exploration and reasoning over the question space and scope. However, such methods struggle with generating outputs that are faithful to the intermediate chain of reasoning produced by the model. On the other end of the spectrum, neuro-symbolic methods such as Faithful CoT (F-CoT) propose to combine LLMs with external symbolic solvers. While such approaches boast a high degree of faithfulness, they usually require a model trained for code generation and struggle with tasks that are ambiguous or hard to formalise strictly. We introduce $ extbf{F}$aithful $ extbf{L}$ogic-$ extbf{A}$ided $ extbf{R}$easoning and $ extbf{E}$xploration ($ extbf{FLARE}$), a novel interpretable approach for traversing the problem space using task decompositions. We use the LLM to plan a solution, soft-formalise the query into facts and predicates using a logic programming code and simulate that code execution using an exhaustive multi-hop search over the defined space. Our method allows us to compute the faithfulness of the reasoning process w.r.t. the generated code and analyse the steps of the multi-hop search without relying on external solvers. Our methods achieve SOTA results on $mathbf{7}$ out of $mathbf{9}$ diverse reasoning benchmarks. We also show that model faithfulness positively correlates with overall performance and further demonstrate that $ extbf{FLARE}$ allows pinpointing the decisive factors sufficient for and leading to the correct answer with optimal reasoning during the multi-hop search.

Problem

• Language Models
• Faithful Reasoning Chains
• Complex Problem Solving

Innovation

• FLARE
• Logical Programming
• Enhanced Reasoning Chain