MA-DAR: Manifold-Aligned Dynamic Adaptive Routing for Continual Temporal Knowledge Graph Reasoning

2026-07-23

AI Summary

This work addresses the norm dominance and semantic ambiguity arising from direct fusion of historical and current representations in continual temporal knowledge graph (TKG) reasoning. To mitigate these issues, the authors propose a lightweight, plug-and-play framework that aligns representations across time steps via manifold alignment to reduce distributional shift, employs a dynamic gating mechanism for adaptive fusion, and incorporates polarization regularization to suppress representational conflicts. Additionally, a replay-based continual learning strategy is integrated to preserve long-term knowledge. Extensive experiments on four public continual TKG benchmarks demonstrate that the proposed framework consistently enhances the performance of mainstream encoders and remains robust across various replay settings, thereby validating its effectiveness and generalizability.

Abstract

Continual temporal knowledge graph (TKG) reasoning aims to continuously incorporate newly emerging facts while preserving previously acquired knowledge. Replay-based continual learning has achieved promising performance by revisiting historical representations. However, existing methods primarily focus on what to replay, while largely overlooking how replayed representations should be integrated with current ones. Such direct integration often gives rise to two critical forms of representation conflict: \textit{norm domination} and \textit{semantic blurring}, ultimately degrading continual reasoning performance. To address these challenges, we propose MA-DAR (Manifold-Aligned Dynamic Adaptive Routing), a lightweight plug-and-play framework for replay representation fusion. MA-DAR first aligns replayed and current representations onto a shared manifold to alleviate distribution discrepancies. It then employs a dynamic gating mechanism to learn dimension-wise fusion weights, adaptively determining the contribution of replayed and current representations to the fused representation. Furthermore, a polarization regularizer encourages more decisive routing behaviors by discouraging ambiguous gating decisions, resulting in more stable and effective knowledge integration. Extensive experiments on four public continual TKG benchmarks demonstrate that MA-DAR consistently improves the performance of representative TKG encoders while remaining effective under different replay settings. Comprehensive ablation studies and visualization analyses further verify the effectiveness of manifold alignment and dynamic adaptive routing in mitigating representation conflicts and improving continual reasoning.

Problem

• continual temporal knowledge graph reasoning
• representation conflict
• norm domination
• semantic blurring
• replay-based continual learning

Innovation

• manifold alignment
• dynamic adaptive routing
• continual learning
• temporal knowledge graph
• representation fusion