Federated Learning of Spiking Neural Networks under Heterogeneous Temporal Resolutions

Sanja Karilanova, Subhrakanti Dey, Ayça Özçelikkale · 2026

Read Federated Learning of Spiking Neural Networks under Heterogeneous Temporal Resolutions by Sanja Karilanova, Subhrakanti Dey, Ayça Özçelikkale (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-2605-15355.txt.

Description

Sanja Karilanova, Subhrakanti Dey, and Ayça Özçelikkale address federated learning of spiking neural networks when edge devices sample at different temporal resolutions. Parameters learned at one rate need not transfer to another, so naive shared averaging falters. Their framework studies how neuron parameters from other rates should be merged at aggregation. Across the SHD and DVS-Gesture benchmarks under varied heterogeneity, the new adaptation methods a lot recover accuracy lost to time mismatch, letting each client train at its local rate.

Catalog record

CREATORSanja Karilanova, Subhrakanti Dey, Ayça Özçelikkale
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE36387
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.15355

Browse the full library · About the Tower of Babel