[2507.08704] Knowledge Fusion via Bidirectional Information Aggregation

Computer Science > Computation and Language arXiv:2507.08704 (cs) [Submitted on 11 Jul 2025 (v1), last revised 23 Mar 2026 (this version, v3)] Title:Knowledge Fusion via Bidirectional Information Aggregation Authors:Songlin Zhai, Guilin Qi, Yue Wang, Yuan Meng View a PDF of the paper titled Knowledge Fusion via Bidirectional Information Aggregation, by Songlin Zhai and 3 other authors View PDF HTML (experimental) Abstract:Knowledge graphs (KGs) are the cornerstone of the semantic web, offering up-to-date representations of real-world entities and relations. Yet large language models (LLMs) remain largely static after pre-training, causing their internal knowledge to become outdated and limiting their utility in time-sensitive web applications. To bridge this gap between dynamic knowledge and static models, a prevalent approach is to enhance LLMs with KGs. However, prevailing methods typically rely on parameter-invasive fine-tuning, which risks catastrophic forgetting and often degrades LLMs' general capabilities. Moreover, their static integration frameworks cannot keep pace with the continuous evolution of real-world KGs, hindering their deployment in dynamic web environments. To bridge this gap, we introduce KGA (\textit{\underline{K}nowledge \underline{G}raph-guided \underline{A}ttention}), a novel framework that dynamically integrates external KGs into LLMs exclusively at inference-time without any parameter modification. Inspired by research on neuroscience, we rewire...