PASCAL: P recise and Efficient A NN- S NN C onversion using Spike Accumulation and A daptive L ayerwise Activation
Pranav Ramesh [email protected], Gopalakrishnan Srinivasan [email protected] · 2025
Read PASCAL: P recise and Efficient A NN- S NN C onversion using Spike Accumulation and A daptive L ayerwise Activation by Pranav Ramesh [email protected], Gopalakrishnan Srinivasan [email protected] (2025) 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-2505-01730.txt.
Description
Pranav Ramesh and Gopalakrishnan Srinivasan present PASCAL, a method for turning standard neural networks into spiking ones, in this 2025 paper. Spiking networks save energy by using sparse adding steps instead of heavy multiply steps. Conversion is a popular way to build deep spiking networks that stay accurate.
Even so, converted networks often need many time steps to match the source network. That delay eats into the energy gains on real data. PASCAL uses spike accumulation and layer-wise adaptive activation to cut the loss. The design aims for accuracy close to the original with fewer steps. Such gains matter for low-power chips in phones and sensors.
Catalog record
| CREATOR | Pranav Ramesh [email protected], Gopalakrishnan Srinivasan [email protected] |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 63 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-01730