STK-Adapter: Incorporating Evolving Graph and Event Chain for Temporal Knowledge Graph Extrapolation

2026-04-20

AI Summary

This work addresses the challenges of temporal knowledge graph (TKG) extrapolation, where shallow alignment between graph structures and large language models leads to loss of spatiotemporal information and dilution of structural features during fine-tuning. To overcome these limitations, the authors propose the STK-Adapter framework, which leverages a mixture-of-experts (MoE) architecture comprising three specialized modules: a spatiotemporal MoE for modeling dynamic graph structures, an event-aware MoE for capturing event chain semantics, and a cross-modal alignment MoE for enabling deep alignment between textual and graphical representations. A TKG-guided attention mechanism is further introduced to facilitate efficient fusion of these components. Experimental results demonstrate that the proposed method significantly outperforms state-of-the-art models across multiple benchmark datasets, exhibiting strong cross-dataset generalization and superior temporal reasoning capabilities.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Temporal Knowledge Graph (TKG) extrapolation aims to predict future events based on historical facts. Recent studies have attempted to enhance TKG extrapolation by integrating TKG's evolving structural representations and textual event chains into Large Language Models (LLMs). Yet, two main challenges limit these approaches: (1) The loss of essential spatial-temporal information due to shallow alignment between TKG's graph evolving structural representation and the LLM's semantic space, and (2) the progressive dilution of the TKG's evolving structural features during LLM fine-tuning. To address these challenges, we propose the Spatial-Temporal Knowledge Adapter (STK-Adapter), which flexibly integrates the evolving graph encoder and the LLM to facilitate TKG reasoning. In STK-Adapter, a Spatial-Temporal MoE is designed to capture spatial structures and temporal patterns inherent in TKGs. An Event-Aware MoE is employed to model intricate temporal semantics dependencies within event chains. In addition, a Cross-Modality Alignment MoE is proposed to facilitate deep cross-modality alignment by TKG-guided attention experts. Extensive experiments on benchmark datasets demonstrate that STK-Adapter significantly outperforms state-of-the-art methods and exhibits strong generalization capabilities in cross-dataset task. The code is available at https://github.com/Zhaoshuyuan0246/STK-Adapter.

Problem

• Temporal Knowledge Graph
• Extrapolation
• Spatial-Temporal Information
• Structural Representation
• Large Language Models

Innovation

• Temporal Knowledge Graph
• Large Language Models
• Mixture of Experts
• Cross-Modality Alignment
• Graph Representation Learning