SAGE: Surrogate-gradient Adaptation via Attention-Guided Entropy for Spiking Transformers

Researchers introduced SAGE, an uncertainty-modulated surrogate-gradient mechanism for Transformer-based spiking neural networks (SNNs), which adapts the surrogate-gradient slope during training using normalized self-attention entropy without changing the inference model. Experiments on CIFAR-10/100 showed consistent accuracy gains of 1-2% over fixed-surrogate baselines across multiple simulation time steps, highlighting attention-derived uncertainty as a lightweight training signal for adaptive surrogate-gradient learning in transformer-based SNNs.

arXiv:2608.13702v1 Announce Type: new Abstract: Spiking neural networks SNNs offer an energy-efficient alternative to conventional deep neural networks by exploiting sparse event-driven computation, but their training remains challenging because the non-differentiable spike function requires surrogate gradients whose fixed shape may be suboptimal across layers and training stages. In this work, we introduce SAGE, an uncertainty-modulated surrogate-gradient mechanism for Transformer-based SNNs. SAGE estimates block-level uncertainty from normalized self-attention entropy and uses this signal to adapt the surrogate-gradient slope during training while leaving the inference model unchanged. By modulating only the training-time surrogate parameter, the proposed method preserves the original architecture and deployment cost while improving optimization flexibility. Experiments on CIFAR-10/100 demonstrate that SAGE achieves improved accuracy over fixed-surrogate baselines, with results up to 1-2\% consistent gains across multiple simulation time steps. These results highlight the potential of attention-derived uncertainty as a lightweight training signal for adaptive surrogate-gradient learning in transformer-based SNNs.