MixCon: Adjusting the Separability of Data Representations for Harder Data Recovery

10/22/2020
by   Xiaoxiao Li, et al.
4

To address the issue that deep neural networks (DNNs) are vulnerable to model inversion attacks, we design an objective function, which adjusts the separability of the hidden data representations, as a way to control the trade-off between data utility and vulnerability to inversion attacks. Our method is motivated by the theoretical insights of data separability in neural networking training and results on the hardness of model inversion. Empirically, by adjusting the separability of data representation, we show that there exist sweet-spots for data separability such that it is difficult to recover data during inference while maintaining data utility.

READ FULL TEXT

page 10

page 25

page 26

research
05/08/2020

Defending Model Inversion and Membership Inference Attacks via Prediction Purification

Neural networks are susceptible to data inference attacks such as the mo...
research
07/13/2021

Deep Neural Networks are Surprisingly Reversible: A Baseline for Zero-Shot Inversion

Understanding the behavior and vulnerability of pre-trained deep neural ...
research
01/26/2022

Variational Model Inversion Attacks

Given the ubiquity of deep neural networks, it is important that these m...
research
03/03/2022

Label-Only Model Inversion Attacks via Boundary Repulsion

Recent studies show that the state-of-the-art deep neural networks are v...
research
06/11/2022

Bilateral Dependency Optimization: Defending Against Model-inversion Attacks

Through using only a well-trained classifier, model-inversion (MI) attac...
research
06/12/2020

Analysis, Design, and Generalization of Electrochemical Impedance Spectroscopy (EIS) Inversion Algorithms

We introduce a framework for analyzing and designing EIS inversion algor...
research
09/09/2020

Analysis of Seismic Inversion with Optimal Transportation and Softplus Encoding

This paper is devoted to theoretical and numerical investigation of the ...

Please sign up or login with your details

Forgot password? Click here to reset