Peeking inside the Black-Box: Reinforcement Learning for Explainable and Accurate Relation Extraction

2025-10-07

AI Summary

Traditional relation extraction (RE) methods suffer from insufficient explanatory supervision, particularly in few-shot settings—exhibiting scattered attention, inadequate keyword capture, and poor interpretability. To address these issues, we propose CogRE, the first RE framework to integrate stepwise reasoning mechanisms inspired by cognitive science. CogRE leverages a large language model–constructed relational lexicon and structured reasoning path generation to extract salient key phrases. We further design a novel reinforcement learning reward function that jointly optimizes both extraction accuracy and explanation quality, enabling end-to-end training. Experiments on the one-shot NYT29 dataset show that CogRE achieves an F1 score of 24.65%, with an absolute improvement of 23.46% after RL optimization. Human evaluation confirms a 54% relative gain in explanation quality. This work establishes a cognitive-inspired paradigm for RE, significantly enhancing model discriminative capability and decision transparency.

Technology Category

• Natural Language Processing: Information Extraction
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Knowledge Representation and Reasoning: Common-Sense Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

This paper introduces a framework for relation extraction (RE) that enhances both accuracy and explainability. The framework has two key components: (i) a reasoning mechanism that formulates relation extraction as a series of text-processing steps inspired by cognitive science, and (ii) an optimization process driven by reinforcement learning (RL) with a novel reward function designed to improve both task accuracy and explanation quality. We call our approach CogRE. Our framework addresses the lack of supervision for language-based explanations in traditional RE by promoting outputs that include important relation keywords. These keywords are drawn from a high-quality dictionary that is automatically constructed using an LLM. We evaluate our approach for the task of one-shot RE using two LLMs and two RE datasets. Our experiments show that CogRE improves explanation quality by addressing two common failure patterns in one-shot RE: poor attention focus and limited one-shot learning capability. For example, our cognitive-structured reasoning with Qwen2.5-15B-Instruct on One-shot NYT29 achieves 24.65% F1, surpassing prior reasoning-based designs. Optimizing this approach with RL using our reward further improves performance by +23.46% (absolute). Finally, human evaluation shows that our best model generates relational keywords closely aligned with gold labels, increasing human explanation quality ratings by 54% (relative).

Problem

• Enhancing relation extraction accuracy and explainability through cognitive reasoning
• Addressing poor attention focus in one-shot relation extraction tasks
• Improving limited one-shot learning capability with reinforcement learning optimization

Innovation

• Reinforcement learning optimizes relation extraction accuracy
• Cognitive science steps structure text processing reasoning
• LLM-built dictionary provides keywords for explanations