Saved in:
Bibliographic Details
Main Authors: Rakhlin, Vlad, Jevnisek, Amir, Avidan, Shai
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2511.11781
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • ReLU activations are the main bottleneck in Private Inference that is based on ResNet networks. This is because they incur significant inference latency. Reducing ReLU count is a discrete optimization problem, and there are two common ways to approach it. Most current state-of-the-art methods are based on a smooth approximation that jointly optimizes network accuracy and ReLU budget at once. However, the last hard thresholding step of the optimization usually introduces a large performance loss. We take an alternative approach that works directly in the discrete domain by leveraging Coordinate Descent as our optimization framework. In contrast to previous methods, this yields a sparse solution by design. We demonstrate, through extensive experiments, that our method is State of the Art on common benchmarks.