SpikeMatch: Semi-Supervised Learning with Temporal Dynamics of Spiking Neural Networks
This addresses the problem of limited labeled data for spiking neural networks, which are energy-efficient but underexplored in semi-supervised learning, representing an incremental advancement in the field.
The paper tackles the lack of semi-supervised learning methods for spiking neural networks by introducing SpikeMatch, a framework that uses temporal dynamics and pseudo-labeling to improve performance with limited labels, achieving better results than adapted methods on standard benchmarks.
Spiking neural networks (SNNs) have recently been attracting significant attention for their biological plausibility and energy efficiency, but semi-supervised learning (SSL) methods for SNN-based models remain underexplored compared to those for artificial neural networks (ANNs). In this paper, we introduce SpikeMatch, the first SSL framework for SNNs that leverages the temporal dynamics through the leakage factor of SNNs for diverse pseudo-labeling within a co-training framework. By utilizing agreement among multiple predictions from a single SNN, SpikeMatch generates reliable pseudo-labels from weakly-augmented unlabeled samples to train on strongly-augmented ones, effectively mitigating confirmation bias by capturing discriminative features with limited labels. Experiments show that SpikeMatch outperforms existing SSL methods adapted to SNN backbones across various standard benchmarks.