BadSNN: Backdoor Attacks on Spiking Neural Networks via Adversarial Spiking Neuron DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY
Abdullah Arafat Miah, Kevin Vu, Yu Bi · 2026
Read BadSNN: Backdoor Attacks on Spiking Neural Networks via Adversarial Spiking Neuron DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY by Abdullah Arafat Miah, Kevin Vu, Yu Bi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-07200.txt.
Description
Spiking neural networks run efficiently by sending information as timed spikes, but their security is less studied than that of usual networks. Abdullah Arafat Miah and two co-authors show how an attacker can exploit the spiking neuron itself, through its hyperparameters, to plant a backdoor.
Their BadSNN attack varies quantities like the membrane threshold and time constant to inject malicious action during training, forcing the model to respond to attacker-chosen triggers. The work matters because energy-efficient spiking hardware is heading for real deployments, and its unique components create unique attack surface. Defenders get an early warning about a threat class that standard backdoor defenses may miss.
Catalog record
| CREATOR | Abdullah Arafat Miah, Kevin Vu, Yu Bi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07200