Information Theoretic Evaluation of Privacy-Leakage, Interpretability, and Transferability for Trustworthy AI

TL;DR

This paper introduces an information theoretic framework for evaluating privacy, interpretability, and transferability in trustworthy AI, using variational models and experiments on benchmark and biomedical data.

Contribution

It presents a unified analytical approach to quantify and optimize the tradeoffs among privacy, interpretability, and transferability in AI models.

Findings

Effective approximation of information measures via variational optimization.

Successful application to benchmark datasets and biomedical stress detection.

Demonstrated tradeoff analysis enhances trustworthy AI development.

Abstract

In order to develop machine learning and deep learning models that take into account the guidelines and principles of trustworthy AI, a novel information theoretic trustworthy AI framework is introduced. A unified approach to "privacy-preserving interpretable and transferable learning" is considered for studying and optimizing the tradeoffs between privacy, interpretability, and transferability aspects. A variational membership-mapping Bayesian model is used for the analytical approximations of the defined information theoretic measures for privacy-leakage, interpretability, and transferability. The approach consists of approximating the information theoretic measures via maximizing a lower-bound using variational optimization. The study presents a unified information theoretic approach to study different aspects of trustworthy AI in a rigorous analytical manner. The approach is demonstrated through numerous experiments on benchmark datasets and a real-world biomedical application concerned with the detection of mental stress on individuals using heart rate variability analysis.