General Self-Prediction Enhancement for Spiking Neurons

2026-01-29

AI Summary

Training spiking neural networks remains challenging due to the non-differentiability of spikes and the inherent trade-offs among performance, efficiency, and biological plausibility; moreover, existing approaches largely overlook the predictive coding mechanisms observed in the brain. Inspired by this, this work introduces predictive coding into spiking neuron design for the first time, proposing self-predictive augmented spiking neurons that generate an internal predictive current from input–output history to modulate membrane potential, thereby establishing a continuous gradient pathway. This mechanism alleviates gradient vanishing, enhances training stability and accuracy, and aligns with biological plasticity principles. Consistent performance improvements are demonstrated across diverse network architectures, neuron types, time steps, and tasks, underscoring the method’s broad applicability and effectiveness.

Technology Category

• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Machine Learning: Neuro-Symbolic Learning
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Spiking Neural Networks (SNNs) are highly energy-efficient due to event-driven, sparse computation, but their training is challenged by spike non-differentiability and trade-offs among performance, efficiency, and biological plausibility. Crucially, mainstream SNNs ignore predictive coding, a core cortical mechanism where the brain predicts inputs and encodes errors for efficient perception. Inspired by this, we propose a self-prediction enhanced spiking neuron method that generates an internal prediction current from its input-output history to modulate membrane potential. This design offers dual advantages, it creates a continuous gradient path that alleviates vanishing gradients and boosts training stability and accuracy, while also aligning with biological principles, which resembles distal dendritic modulation and error-driven synaptic plasticity. Experiments show consistent performance gains across diverse architectures, neuron types, time steps, and tasks demonstrating broad applicability for enhancing SNNs.

Problem

• Spiking Neural Networks
• spike non-differentiability
• predictive coding
• training stability
• biological plausibility

Innovation

• Self-Prediction
• Spiking Neural Networks
• Predictive Coding
• Gradient Stabilization
• Biological Plausibility