Bridge lp improves link prediction by uniting text and graph data
Bridging Semantics and Structure: A Theoretically Grounded Framework for Link Prediction on TAGs
Artificial Intelligence
Summary
Predicting missing links in graphs that have text descriptions is important for tasks like recommending items or discovering knowledge. The authors present Bridge-LP, a new method that smartly combines graph and language information using three key techniques to better learn from the structure and text. Their method is both simple and efficient, scaling to large datasets. Experiments show Bridge-LP outperforms many existing methods by a large margin across multiple test cases.
What this means in practice
- •For recommendation system teams: Improve recommendation accuracy by better predicting links using combined language and graph data representations.
- •For knowledge graph engineers: Enhance knowledge discovery by more effectively predicting missing relationships in large text-enriched graphs.
Authors
Chen Shao, Ruirui Wang, Tobias Kafer, Danai Koutra
Abstract
Link prediction on text-attributed graphs (TAGs) is a fundamental problem with applications in knowledge discovery, recommendation and beyond. Recent progress has highlighted the promise of combining Graph Neural Networks (GNNs) with Language Models (LMs) in order to leverage the complementary structural and semantic signals. However, due to the distinct inductive biases of GNNs and LMs, the design of an optimal fusion strategy remains an open research question. In this work, we present Bridge-LP, an intuitive, simple, efficient and theoretically grounded framework that unifies graph and language representations. It is built around three theoretically-justified design components: (D1) jumping connection of textual proximity, which enables robust LM training; (D2) compressed embedding, which encodes neighborhood information long exploited by classical link prediction heuristics, thereby enriching structural signals in a learnable way; and (D3) cache embedding training strategy, which reduces computational complexity. Together, these mechanisms establish an end-to-end cross-modal framework that scales to large graphs while capturing richer graph-text interactions. We conduct an extensive empirical evaluation, comparing against 40 existing models across 9 diverse datasets. Experiments show that Bridge-LP consistently outperforms state-of-the-art approaches, delivering up to a 43.06% improvement in Mean Reciprocal Rank (MRR) and setting a new benchmark for language-graph integration.