SpikeMatch: Semi-Supervised Learning with Temporal Dynamics of Spiking Neural Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yang, Jini, Oh, Beomseok, Kim, Seungryong, Kim, Sunok
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911178721067008
author Yang, Jini
Oh, Beomseok
Kim, Seungryong
Kim, Sunok
author_facet Yang, Jini
Oh, Beomseok
Kim, Seungryong
Kim, Sunok
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2509_22581
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SpikeMatch: Semi-Supervised Learning with Temporal Dynamics of Spiking Neural Networks
Yang, Jini
Oh, Beomseok
Kim, Seungryong
Kim, Sunok
Computer Vision and Pattern Recognition
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.
title SpikeMatch: Semi-Supervised Learning with Temporal Dynamics of Spiking Neural Networks
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.22581