ChronoSpike: An Adaptive Spiking Graph Neural Network for Dynamic Graphs

TL;DR

ChronoSpike tackles dynamic graph representation learning by fusing adaptive spiking neurons, a multi-head attentive spatial encoder on continuous features, and a Transformer-based temporal encoder. It achieves linear memory in the temporal horizon while maintaining expressive power, backed by stability guarantees for membrane dynamics and gradient flow. Empirically, it outperforms a wide range of baselines on large-scale temporal node classification and exhibits strong robustness, interpretability, and efficiency. The combination of adaptive per-channel spike dynamics, global temporal aggregation, and contrastive regularization offers a practical, neuromorphic-friendly approach to evolving graphs with real-world impact.

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 $ρ< 1$, and BIBO stability; interpretability analyses reveal heterogeneous temporal receptive fields and a learned primacy effect with $83-88\%$ sparsity.

Figures (9)

• Figure 1: Overview of the ChronoSpike framework. The dynamic graph is represented as a sequence of snapshots. At each time step, node features are aggregated from sampled neighborhoods using a multi-head attentive spatial aggregator and encoded into spike signals via adaptive LIF neurons. Temporal dependencies across snapshots are captured by a lightweight Transformer-based temporal aggregation module with learnable positional encodings. The final node representations are used for downstream prediction.
• Figure 2: Overhead comparison of different methods in terms of model parameter size and average training time per epoch. Models that do not scale to the Patent dataset or do not report overhead statistics are excluded from the comparison.
• Figure 3: Parameter sensitivity analysis on DBLP (left) and Tmall (right) datasets (80% training). (a) Learning rate vs. dropout rate heatmap showing optimal regions. (b) SNN parameter $\alpha$ vs. contrastive weight showing robust performance. Darker colors indicate higher Micro-F1 scores.
• Figure 4: Sensitivity of ChronoSpike to (a) sampling probability, (b) batch size, and (c) hidden dimension size on DBLP (left) and Tmall (right) datasets using an 80% training split.
• Figure 5: Training loss curves of ChronoSpike on DBLP, Tmall, and Patent with 40%, 60%, and 80% training splits. ChronoSpike shows stable convergence across all settings, with faster loss reduction at higher supervision.

Theorems & Definitions (13)

• Theorem 4.1: Boundedness of Adaptive LIF Dynamics
• proof
• Theorem 4.2: Network-Level Boundedness and BIBO Stability
• proof
• Theorem 4.3: Gradient Flow Stability and Conditional Convergence
• proof
• Theorem 4.4: Expressiveness in the WL Hierarchy (Static Graphs)
• proof
• Theorem 4.5: Generalization Bound with Sparse Spiking
• proof : Proof Sketch