Attention Dispersion in Dynamic Graph Transformers: Diagnosis and a Transferable Fix

2026-05-15

AI Summary

This work addresses a common failure mode in dynamic graph Transformers under temporal distribution shifts, where attention dispersion hinders the model’s ability to focus on critical nodes, leading to performance degradation. The study is the first to diagnose this issue and introduces a transferable differential attention mechanism that mitigates the problem by amplifying attention weights for nodes exhibiting strong predictive signals. Building upon this insight, the proposed DiffDyG model integrates continuous-time dynamic graph modeling with standard input encodings. It achieves state-of-the-art performance across nine benchmark datasets and three negative sampling protocols, demonstrating particularly significant gains in scenarios with pronounced temporal distribution shifts.

Abstract

Transformer-based architectures have become the dominant paradigm for Continuous-Time Dynamic Graph (CTDG) learning, yet their performance remains limited on temporally shifted datasets. In this work, we identify attention dispersion as a shared failure mode of dynamic graph Transformers under temporal distribution shift. Through controlled ablation contrasting structurally and temporally distinguished historical neighbors against random ones, we show that prediction depends on a class of critical nodes that carry consistently more predictive signal than arbitrary neighbors. However, existing Transformers fail to focus on these nodes even when they are present in the input, as temporal shift weakens attention contrast and produces overly dispersed attention distributions. This diagnosis suggests a simple and transferable fix: replace standard attention with differential attention, which suppresses common-mode attention and amplifies distinctive token-level signals. When added to three representative CTDG Transformer baselines, differential attention consistently improves performance, with gains concentrated on high-shift datasets. Attention-level measurements further confirm the mechanism, showing reduced attention entropy and increased attention mass on critical nodes. Building on these findings, we introduce DiffDyG, a reference implementation combining differential attention with standard input encodings. Across 9 benchmarks and three negative sampling protocols, DiffDyG achieves SOTA performance, with especially large gains on the most shifted datasets.

Problem

• attention dispersion
• temporal distribution shift
• dynamic graph transformers
• critical nodes
• continuous-time dynamic graph

Innovation

• attention dispersion
• differential attention
• temporal distribution shift
• dynamic graph transformers
• critical nodes