Selective Network Linearization for Efficient Private Inference

02/04/2022
by   Minsu Cho, et al.
1

Private inference (PI) enables inference directly on cryptographically secure data. While promising to address many privacy issues, it has seen limited use due to extreme runtimes. Unlike plaintext inference, where latency is dominated by FLOPs, in PI non-linear functions (namely ReLU) are the bottleneck. Thus, practical PI demands novel ReLU-aware optimizations. To reduce PI latency we propose a gradient-based algorithm that selectively linearizes ReLUs while maintaining prediction accuracy. We evaluate our algorithm on several standard PI benchmarks. The results demonstrate up to 4.25% more accuracy (iso-ReLU count at 50K) or 2.2× less latency (iso-accuracy at 70%) than the current state of the art and advance the Pareto frontier across the latency-accuracy space. To complement empirical results, we present a "no free lunch" theorem that sheds light on how and when network linearization is possible while maintaining prediction accuracy.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
03/02/2021

DeepReDuce: ReLU Reduction for Fast Private Inference

The recent rise of privacy concerns has led researchers to devise method...
research
08/20/2023

AutoReP: Automatic ReLU Replacement for Fast Private Network Inference

The growth of the Machine-Learning-As-A-Service (MLaaS) market has highl...
research
06/15/2020

CryptoNAS: Private Inference on a ReLU Budget

Machine learning as a service has given raise to privacy concerns surrou...
research
06/17/2021

Sphynx: ReLU-Efficient Network Design for Private Inference

The emergence of deep learning has been accompanied by privacy concerns ...
research
06/15/2021

Circa: Stochastic ReLUs for Private Deep Learning

The simultaneous rise of machine learning as a service and concerns over...
research
10/22/2020

CryptoGRU: Low Latency Privacy-Preserving Text Analysis With GRU

Billions of text analysis requests containing private emails, personal t...
research
04/20/2023

DeepReShape: Redesigning Neural Networks for Efficient Private Inference

The increasing demand for privacy and security has driven the advancemen...

Please sign up or login with your details

Forgot password? Click here to reset