On the Subadditivity of Mutual f-Information
Abstract
We address an open problem raised by Polyanskiy and Wu in their textbook concerning the subadditivity of mutual f-information. We study two notions of generalized mutual information: mutual f-information, defined by Ĩf(X;Y)=Df(pXY|pXpY), and variational (G,f)-information, obtained by minimizing an average of transformed f-divergences. For the first notion, we derive a sufficient condition on g(x)=x2f''(x) that guarantees subadditivity under conditionally independent observations. This condition yields explicit families beyond the previously known examples generated by xlog x and (x−1)log x. Our proof relies on the concavity of a generalized variance functional. For the variational notion, we extend existing sufficient conditions beyond the transforms G(x)∈{x,log(1+x),−log(1−x)}. Using a comparison inequality for G−1(G(A)+G(B)), we obtain differential criteria for concave and convex transforms. Explicit constructions illustrate the resulting classes of subadditive information measures.
Related
- Free Mutual Information and Higher-Point Out-of-Time-Ordered Correlators
- Mutual Information in High-Dimensional Neural Coding
- Dynamical evolution of quantum mutual information in Schwarzschild spacetime
- Mutual Information of Measurement in Conformal Field Theories
- Correlation Harvesting by Unruh--DeWitt Detectors in a Cylindrical Cavity: Entanglement, Mutual Information, and Quantum Discord
BibTeX
@misc{abin2026subadditivity,
title = {{On the Subadditivity of Mutual f-Information}},
author = {Hamidreza Abin and Amin Gohari},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.21231},
eprint = {2610.21231},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.21231},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - On the Subadditivity of Mutual f-Information
AU - Hamidreza Abin
AU - Amin Gohari
PY - 2026
DA - 2026/10/15
PB - arXiv
DO - 10.48550/arxiv.2610.21231
UR - https://arxiv.org/abs/2610.21231
AB - We address an open problem raised by Polyanskiy and Wu in their textbook concerning the subadditivity of mutual f-information. We study two notions of generalized mutual information: mutual f-information, defined by Ĩ_f(X;Y)=D_f(p_XY|p_Xp_Y), and variational (G,f)-information, obtained by minimizing an average of transformed f-divergences. For the first notion, we derive a sufficient condition on g(x)=x^2f''(x) that guarantees subadditivity under conditionally independent observations. This condition yields explicit families beyond the previously known examples generated by xlog x and (x−1)log x. Our proof relies on the concavity of a generalized variance functional. For the variational notion, we extend existing sufficient conditions beyond the transforms G(x)∈{x,log(1+x),−log(1−x)}. Using a comparison inequality for G^−1(G(A)+G(B)), we obtain differential criteria for concave and convex transforms. Explicit constructions illustrate the resulting classes of subadditive information measures.
ER -