[2505.05375] Threshold Modulation for Online Test-Time Adaptation of Spiking Neural Networks

Computer Science > Computer Vision and Pattern Recognition arXiv:2505.05375 (cs) [Submitted on 8 May 2025 (v1), last revised 4 Apr 2026 (this version, v3)] Title:Threshold Modulation for Online Test-Time Adaptation of Spiking Neural Networks Authors:Kejie Zhao, Wenjia Hua, Aiersi Tuerhong, Luziwei Leng, Yuxin Ma, Qinghai Guo View a PDF of the paper titled Threshold Modulation for Online Test-Time Adaptation of Spiking Neural Networks, by Kejie Zhao and 5 other authors View PDF HTML (experimental) Abstract:Recently, spiking neural networks (SNNs), deployed on neuromorphic chips, provide highly efficient solutions on edge devices in different scenarios. However, their ability to adapt to distribution shifts after deployment has become a crucial challenge. Online test-time adaptation (OTTA) offers a promising solution by enabling models to dynamically adjust to new data distributions without requiring source data or labeled target samples. Nevertheless, existing OTTA methods are largely designed for traditional artificial neural networks and are not well-suited for SNNs. To address this gap, we propose a low-power, neuromorphic chip-friendly online test-time adaptation framework, aiming to enhance model generalization under distribution shifts. The proposed approach is called Threshold Modulation (TM), which dynamically adjusts the firing threshold through neuronal dynamics-inspired normalization, being more compatible with neuromorphic hardware. Experimental results on bench...