ChronoSpike: An Adaptive Spiking Graph Neural Network for Dynamic Graphs

2026-02-01

AI Summary

This work addresses the challenge in dynamic graph representation learning of simultaneously modeling structural relationships and temporal evolution, where existing methods often suffer from limited computational efficiency, gradient instability, or inadequate global context capture. To this end, we propose ChronoSpike, the first framework to integrate adaptive spiking neural networks into dynamic graph learning. ChronoSpike combines learnable leaky integrate-and-fire (LIF) neurons, channel-wise membrane potential dynamics, multi-head attention for spatial aggregation, and a lightweight Transformer-based temporal encoder. This design enables fine-grained local modeling and long-range dependency capture while maintaining linear memory complexity and a fixed parameter count of only 105K. Experiments on three large-scale benchmarks show that ChronoSpike improves Macro-F1 and Micro-F1 by 2.0% and 2.4%, respectively, trains 3–10× faster than recurrent approaches, and achieves 83–88% activation sparsity with theoretical stability guarantees.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Graph-based Machine Learning
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Dynamic graph representation learning requires capturing both structural relationships and temporal evolution, yet existing approaches face a fundamental trade-off: attention-based methods achieve expressiveness at $O(T^2)$ complexity, while recurrent architectures suffer from gradient pathologies and dense state storage. Spiking neural networks offer event-driven efficiency but remain limited by sequential propagation, binary information loss, and local aggregation that misses global context. We propose ChronoSpike, an adaptive spiking graph neural network that integrates learnable LIF neurons with per-channel membrane dynamics, multi-head attentive spatial aggregation on continuous features, and a lightweight Transformer temporal encoder, enabling both fine-grained local modeling and long-range dependency capture with linear memory complexity $O(T \cdot d)$. On three large-scale benchmarks, ChronoSpike outperforms twelve state-of-the-art baselines by $2.0\%$ Macro-F1 and $2.4\%$ Micro-F1 while achieving $3-10\times$ faster training than recurrent methods with a constant 105K-parameter budget independent of graph size. We provide theoretical guarantees for membrane potential boundedness, gradient flow stability under contraction factor $\rho<1$, and BIBO stability; interpretability analyses reveal heterogeneous temporal receptive fields and a learned primacy effect with $83-88\%$ sparsity.

Problem

• dynamic graph representation learning
• spiking neural networks
• temporal evolution
• graph neural networks
• computational complexity

Innovation

• Spiking Neural Networks
• Dynamic Graph Representation Learning
• Adaptive LIF Neurons
• Linear Memory Complexity
• Temporal Transformer Encoder