An Encryption Method of ConvMixer Models without Performance Degradation

TL;DR

This paper introduces a novel encryption method for ConvMixer models that preserves their performance while enhancing privacy and security, without requiring additional training data or network modifications.

Contribution

The paper presents a new encryption technique for ConvMixer models that maintains accuracy and does not need extra training data or network changes.

Findings

Encrypted models achieve the same accuracy as plain models on CIFAR10.

The method effectively protects models without performance degradation.

No additional training data or network modifications are needed.

Abstract

In this paper, we propose an encryption method for ConvMixer models with a secret key. Encryption methods for DNN models have been studied to achieve adversarial defense, model protection and privacy-preserving image classification. However, the use of conventional encryption methods degrades the performance of models compared with that of plain models. Accordingly, we propose a novel method for encrypting ConvMixer models. The method is carried out on the basis of an embedding architecture that ConvMixer has, and models encrypted with the method can have the same performance as models trained with plain images only when using test images encrypted with a secret key. In addition, the proposed method does not require any specially prepared data for model training or network modification. In an experiment, the effectiveness of the proposed method is evaluated in terms of classification accuracy and model protection in an image classification task on the CIFAR10 dataset.