Loading [MathJax]/extensions/MathMenu.js
Robust Machine Learning via Privacy/ Rate-Distortion Theory | IEEE Conference Publication | IEEE Xplore

Robust Machine Learning via Privacy/ Rate-Distortion Theory


Abstract:

Robust machine learning formulations have emerged to address the prevalent vulnerability of deep neural networks to adversarial examples. Our work draws the connection be...Show More

Abstract:

Robust machine learning formulations have emerged to address the prevalent vulnerability of deep neural networks to adversarial examples. Our work draws the connection between optimal robust learning and the privacy-utility tradeoff problem, which is a generalization of the rate-distortion problem. The saddle point of the game between a robust classifier and an adversarial perturbation can be found via the solution of a maximum conditional entropy problem. This information-theoretic perspective sheds light on the fundamental tradeoff between robustness and clean data performance, which ultimately arises from the geometric structure of the underlying data distribution and perturbation constraints.
Date of Conference: 12-20 July 2021
Date Added to IEEE Xplore: 01 September 2021
ISBN Information:
Conference Location: Melbourne, Australia

References

References is not available for this document.