DeepReDuce: ReLU Reduction for Fast Private Inference

TL;DR

DeepReDuce introduces a method to selectively remove ReLU activations from neural networks to significantly reduce inference latency in privacy-preserving settings while maintaining high accuracy, enabling more practical private inference.

Contribution

It proposes a novel ReLU reduction technique that optimizes the tradeoff between network accuracy and inference speed in private neural inference.

Findings

Reduces ReLU count by up to 3.5 times

Maintains high accuracy with fewer ReLUs

Outperforms state-of-the-art private inference methods

Abstract

The recent rise of privacy concerns has led researchers to devise methods for private neural inference -- where inferences are made directly on encrypted data, never seeing inputs. The primary challenge facing private inference is that computing on encrypted data levies an impractically-high latency penalty, stemming mostly from non-linear operators like ReLU. Enabling practical and private inference requires new optimization methods that minimize network ReLU counts while preserving accuracy. This paper proposes DeepReDuce: a set of optimizations for the judicious removal of ReLUs to reduce private inference latency. The key insight is that not all ReLUs contribute equally to accuracy. We leverage this insight to drop, or remove, ReLUs from classic networks to significantly reduce inference latency and maintain high accuracy. Given a target network, DeepReDuce outputs a Pareto frontier of networks that tradeoff the number of ReLUs and accuracy. Compared to the state-of-the-art for private inference DeepReDuce improves accuracy and reduces ReLU count by up to 3.5% (iso-ReLU count) and 3.5$\times$ (iso-accuracy), respectively.