A Batch-Insensitive Dynamic GNN Approach to Address Temporal Discontinuity in Graph Streams

TL;DR

This paper introduces BADGNN, a batch-agnostic dynamic GNN framework that preserves temporal continuity in graph streams, improves training efficiency, and maintains high performance by controlling parameter search space and attention distortion.

Contribution

The paper proposes BADGNN, a novel approach combining Temporal Lipschitz Regularization and Adaptive Attention Adjustment to address temporal discontinuity in dynamic GNNs.

Findings

BADGNN enables larger batch sizes and faster training.

It maintains strong performance on benchmark datasets.

Theoretical analysis shows reduced parameter search space.

Abstract

In dynamic graphs, preserving temporal continuity is critical. However, Memory-based Dynamic Graph Neural Networks (MDGNNs) trained with large batches often disrupt event sequences, leading to temporal information loss. This discontinuity not only deteriorates temporal modeling but also hinders optimization by increasing the difficulty of parameter convergence. Our theoretical study quantifies this through a Lipschitz upper bound, showing that large batch sizes enlarge the parameter search space. In response, we propose BADGNN, a novel batch-agnostic framework consisting of two core components: (1) Temporal Lipschitz Regularization (TLR) to control parameter search space expansion, and (2) Adaptive Attention Adjustment (A3) to alleviate attention distortion induced by both regularization and batching. Empirical results on three benchmark datasets show that BADGNN maintains strong performance while enabling significantly larger batch sizes and faster training compared to TGN. Our code is available at Code: https://anonymous.4open.science/r/TGN_Lipichitz-C033/.