Metraipolis › archive › dc_70f34877dee5
Activation-Energy Pruning for Spiking Neural Networks: Unsupervised Personalization via Spike-Count Saliency — arXiv:2609.26167
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.26167
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T15:07:48.032Z
- SHA-256 of the archived text
c63678e8204c80c5212f35c6e40269a22c4257057da6f1062f6d0c4b173de22e- Stable URL of this page
https://metraipolis.com/documento/dc_70f34877dee5
How to cite
Activation-Energy Pruning for Spiking Neural Networks: Unsupervised Personalization via Spike-Count Saliency — arXiv:2609.26167 — arXiv (en), https://arxiv.org/abs/2609.26167 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_70f34877dee5 (Metraipolis archive dc_70f34877dee5, sha256 c63678e8204c80c5212f35c6e40269a22c4257057da6f1062f6d0c4b173de22e)
BibTeX
@misc{dc70f34877dee5,
title = {Activation-Energy Pruning for Spiking Neural Networks: Unsupervised Personalization via Spike-Count Saliency — arXiv:2609.26167},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.26167},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_70f34877dee5; sha256 c63678e8204c80c5212f35c6e40269a22c4257057da6f1062f6d0c4b173de22e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.26167. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Activation-energy pruning -- removing weights whose product of magnitude and cumulative pre-synaptic spike count falls below a threshold -- was established as an effective unsupervised personalization strategy for conventional deep neural networks~\citep{BINGHAM2025101242}. This paper asks what happens when the same criterion is applied to spiking neural networks (SNNs), where activation energy is not merely a useful heuristic but a literal physical quantity proportional to the metabolic cost of each synapse. The answer is surprising on three counts. First, gradient-based pruning methods that perform competitively on conventional networks (SNIP, GraSP, magnitude pruning) consistently underperform on SNNs, collapsing to near-chance accuracy by $\sigma = 0.2$ sparsity across all tested architectures and datasets. We trace this to a systematic incompatibility between surrogate-gradient saliency estimation and the binary spike-train representation, though we cannot rule out that alternative surrogate choices or hyperparameter settings might partially mitigate the effect. Second, activation-energy pruning applied to a neuromorphic benchmark \emph{improves} over the source model at high sparsity ($98.4 \pm 0.4\%$ vs.\ $97.2 \pm 0.7\%$ at $\sigma = 0.8$ on N-MNIST), a phenomenon with no counterpart in the conventional network setting. We interpret this result as consistent with experience-dependent cortical specialisation: removing connections active only for non-target classes may reduce cross-class interference and produce a cleaner target representation, though we note this is an interpretive analogy rather than a mechanistic demonstration.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-23 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG) ; Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2609.26167 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T14:37:29.462Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).