On Dependence Measures Based on f f -Divergence and Φ \Phi -Entropy

Chenyu Wang, Amin Gohari · 2025

Read On Dependence Measures Based on f f -Divergence and Φ \Phi -Entropy by Chenyu Wang, Amin Gohari (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-03754.txt.

Description

Chenyu Wang and Amin Gohari build new ways to measure how strongly two random quantities depend on each other, generalizing a classic information measure. Their paper appeared in two thousand twenty-five. Their tools grow from a broad family of divergence measures between distributions. A conditional version follows from an exact chain rule they develop.

That version, though, lacks two tidy properties: symmetry and self-reduction. To restore them, the authors wrap the measure in a variational envelope with helper variables. Both forms collapse to Shannon's mutual information for the classic choice of generator. For smooth generators they pin down the optimizer and a closed form. A striking split emerges for a privacy-assisted version: it is either a simple rescaling of Shannon's measure or infinite. Bounds for network information regions round out the work.

Catalog record

CREATORChenyu Wang, Amin Gohari
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE137 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-03754

Browse the full library · About the Tower of Babel