Bilateral Dependency Optimization: Defending Against Model-inversion Attacks

TL;DR

This paper introduces a bilateral dependency optimization strategy to defend against model-inversion attacks by balancing dependencies between inputs, outputs, and latent representations, achieving state-of-the-art privacy protection with minimal utility loss.

Contribution

It proposes a novel bilateral dependency optimization (BiDO) approach that enhances privacy defense against MI attacks while maintaining classification performance.

Findings

BiDO outperforms existing defenses across various datasets and attacks.

BiDO achieves state-of-the-art privacy protection with minor accuracy loss.

Two implementations, BiDO-COCO and BiDO-HSIC, demonstrate effectiveness.

Abstract

Through using only a well-trained classifier, model-inversion (MI) attacks can recover the data used for training the classifier, leading to the privacy leakage of the training data. To defend against MI attacks, previous work utilizes a unilateral dependency optimization strategy, i.e., minimizing the dependency between inputs (i.e., features) and outputs (i.e., labels) during training the classifier. However, such a minimization process conflicts with minimizing the supervised loss that aims to maximize the dependency between inputs and outputs, causing an explicit trade-off between model robustness against MI attacks and model utility on classification tasks. In this paper, we aim to minimize the dependency between the latent representations and the inputs while maximizing the dependency between latent representations and the outputs, named a bilateral dependency optimization (BiDO) strategy. In particular, we use the dependency constraints as a universally applicable regularizer in addition to commonly used losses for deep neural networks (e.g., cross-entropy), which can be instantiated with appropriate dependency criteria according to different tasks. To verify the efficacy of our strategy, we propose two implementations of BiDO, by using two different dependency measures: BiDO with constrained covariance (BiDO-COCO) and BiDO with Hilbert-Schmidt Independence Criterion (BiDO-HSIC). Experiments show that BiDO achieves the state-of-the-art defense performance for a variety of datasets, classifiers, and MI attacks while suffering a minor classification-accuracy drop compared to the well-trained classifier with no defense, which lights up a novel road to defend against MI attacks.