Revisiting the Variational Information Bottleneck

Nir Zvi Weingarten, Moshe Butman, Zohar Yakhini, Ronit Bustin

Keywords: information bottleneck, information theory, representation learning, adversarial attacks, regularization, supervised learning

TL;DR: A new variational adaptation of the IB for supervised DNN optimization

Abstract: The Information Bottleneck (IB) framework offers a theoretically optimal approach to data modeling, though it is often intractable. Recent efforts have optimized supervised deep neural networks (DNNs) using a variational upper bound on the IB objective, leading to enhanced robustness to adversarial attacks. In these studies, supervision assumes a dual role: sometimes as a presumably constant and observed random variable, and at other times as its variational approximation. This work proposes an extension to the IB framework, and consequently to the derivation of its variational bound, that resolves this duality. Applying the resulting bound as an objective for supervised DNNs induces significant empirical improvements, and provides an information theoretic motivation for decoder regularization.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 7728