GS-Quant: Granular Semantic and Generative Structural Quantization for Knowledge Graph Completion

arXiv cs.AI / 4/25/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces GS-Quant, a framework for knowledge graph completion that bridges continuous graph embeddings and discrete LLM token representations.
Unlike prior quantization methods that compress numbers without preserving meaning, GS-Quant produces discrete codes that are both semantically coherent and structurally stratified.
GS-Quant uses a Granular Semantic Enhancement module to encode coarse-to-fine hierarchical knowledge, where early codes capture global categories and later codes refine detailed attributes.
It also includes a Generative Structural Reconstruction module that imposes causal dependencies across the code sequence, turning independent units into structured semantic descriptors.
Experiments show GS-Quant improves over existing text-based and embedding-based baselines, and the authors make the code publicly available on GitHub.

Abstract

Large Language Models (LLMs) have shown immense potential in Knowledge Graph Completion (KGC), yet bridging the modality gap between continuous graph embeddings and discrete LLM tokens remains a critical challenge. While recent quantization-based approaches attempt to align these modalities, they typically treat quantization as flat numerical compression, resulting in semantically entangled codes that fail to mirror the hierarchical nature of human reasoning. In this paper, we propose GS-Quant, a novel framework that generates semantically coherent and structurally stratified discrete codes for KG entities. Unlike prior methods, GS-Quant is grounded in the insight that entity representations should follow a linguistic coarse-to-fine logic. We introduce a Granular Semantic Enhancement module that injects hierarchical knowledge into the codebook, ensuring that earlier codes capture global semantic categories while later codes refine specific attributes. Furthermore, a Generative Structural Reconstruction module imposes causal dependencies on the code sequence, transforming independent discrete units into structured semantic descriptors. By expanding the LLM vocabulary with these learned codes, we enable the model to reason over graph structures isomorphically to natural language generation. Experimental results demonstrate that GS-Quant significantly outperforms existing text-based and embedding-based baselines. Our code is publicly available at https://github.com/mikumifa/GS-Quant.