DEPTH: Hallucination-Free Relation Extraction via Dependency-Aware Sentence Simplification and Two-tiered Hierarchical Refinement

2025-08-19

AI Summary

Large language models (LLMs) suffer from hallucination in relation extraction—particularly under complex syntactic and semantic conditions—leading to spurious relation predictions that corrupt knowledge graphs and undermine downstream task reliability. To address this, we propose a dependency-aware sentence simplification framework coupled with a two-tier hierarchical refinement strategy: (1) local relation modeling via shortest dependency path extraction, and (2) global consistency correction guided by a causal-aware reward model that integrates reinforcement learning with human feedback to mitigate reward hacking. Our approach significantly enhances robustness and interpretability in relation judgment. Evaluated on six benchmark datasets, it reduces average hallucination rate to 7.0% and improves F1 score by 17.2% over state-of-the-art methods. This work establishes a trustworthy, controllable paradigm for LLM-based relation extraction.

Technology Category

• Natural Language Processing: Information Extraction
• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Relation extraction enables the construction of structured knowledge for many downstream applications. While large language models (LLMs) have shown great promise in this domain, most existing methods concentrate on relation classification, which predicts the semantic relation type between a related entity pair. However, we observe that LLMs often struggle to reliably determine whether a relation exists, especially in cases involving complex sentence structures or intricate semantics, which leads to spurious predictions. Such hallucinations can introduce noisy edges in knowledge graphs, compromising the integrity of structured knowledge and downstream reliability. To address these challenges, we propose DEPTH, a framework that integrates Dependency-aware sEntence simPlification and Two-tiered Hierarchical refinement into the relation extraction pipeline. Given a sentence and its candidate entity pairs, DEPTH operates in two stages: (1) the Grounding module extracts relations for each pair by leveraging their shortest dependency path, distilling the sentence into a minimal yet coherent relational context that reduces syntactic noise while preserving key semantics; (2) the Refinement module aggregates all local predictions and revises them based on a holistic understanding of the sentence, correcting omissions and inconsistencies. We further introduce a causality-driven reward model that mitigates reward hacking by disentangling spurious correlations, enabling robust fine-tuning via reinforcement learning with human feedback. Experiments on six benchmarks demonstrate that DEPTH reduces the average hallucination rate to 7.0% while achieving a 17.2% improvement in average F1 score over state-of-the-art baselines.

Problem

• Reducing relation extraction hallucinations in complex sentences
• Eliminating spurious predictions from LLMs in knowledge graphs
• Addressing syntactic noise and semantic intricacies in RE

Innovation

• Dependency-aware sentence simplification for relational context
• Two-tiered hierarchical refinement of local predictions
• Causality-driven reward model for robust fine-tuning