Geodesic
Properties of unique information
Kybernetika, Tome 57 (2021) no. 3, pp. 383-403 Cet article a éte moissonné depuis la source Czech Digital Mathematics Library

Voir la notice de l'article

We study the unique information function $UI(T:X\setminus Y)$ defined by Bertschinger et al. within the framework of information decompositions. In particular, we study uniqueness and support of the solutions to the convex optimization problem underlying the definition of $UI$. We identify sufficient conditions for non-uniqueness of solutions with full support in terms of conditional independence constraints and in terms of the cardinalities of $T$, $X$ and $Y$. Our results are based on a reformulation of the first order conditions on the objective function as rank constraints on a matrix of conditional probabilities. These results help to speed up the computation of $UI(T:X\setminus Y)$, most notably when $T$ is binary. Optima in the relative interior of the optimization domain are solutions of linear equations if $T$ is binary. In the all binary case, we obtain a complete picture of where the optimizing probability distributions lie.
We study the unique information function $UI(T:X\setminus Y)$ defined by Bertschinger et al. within the framework of information decompositions. In particular, we study uniqueness and support of the solutions to the convex optimization problem underlying the definition of $UI$. We identify sufficient conditions for non-uniqueness of solutions with full support in terms of conditional independence constraints and in terms of the cardinalities of $T$, $X$ and $Y$. Our results are based on a reformulation of the first order conditions on the objective function as rank constraints on a matrix of conditional probabilities. These results help to speed up the computation of $UI(T:X\setminus Y)$, most notably when $T$ is binary. Optima in the relative interior of the optimization domain are solutions of linear equations if $T$ is binary. In the all binary case, we obtain a complete picture of where the optimizing probability distributions lie.
DOI : 10.14736/kyb-2021-3-0383
Classification : 94A15, 94A17
Keywords: information decomposition; unique information 
@article{10_14736_kyb_2021_3_0383,
     author = {Rauh, Johannes and Sch\"unemann, Maik and Jost, J\"urgen},
     title = {Properties of unique information},
     journal = {Kybernetika},
     pages = {383--403},
     year = {2021},
     volume = {57},
     number = {3},
     doi = {10.14736/kyb-2021-3-0383},
     mrnumber = {4299455},
     zbl = {07442516},
     language = {en},
     url = {http://geodesic.mathdoc.fr/articles/10.14736/kyb-2021-3-0383/}
}
TY  - JOUR
AU  - Rauh, Johannes
AU  - Schünemann, Maik
AU  - Jost, Jürgen
TI  - Properties of unique information
JO  - Kybernetika
PY  - 2021
SP  - 383
EP  - 403
VL  - 57
IS  - 3
UR  - http://geodesic.mathdoc.fr/articles/10.14736/kyb-2021-3-0383/
DO  - 10.14736/kyb-2021-3-0383
LA  - en
ID  - 10_14736_kyb_2021_3_0383
ER  - 
%0 Journal Article
%A Rauh, Johannes
%A Schünemann, Maik
%A Jost, Jürgen
%T Properties of unique information
%J Kybernetika
%D 2021
%P 383-403
%V 57
%N 3
%U http://geodesic.mathdoc.fr/articles/10.14736/kyb-2021-3-0383/
%R 10.14736/kyb-2021-3-0383
%G en
%F 10_14736_kyb_2021_3_0383
Rauh, Johannes; Schünemann, Maik; Jost, Jürgen. Properties of unique information. Kybernetika, Tome 57 (2021) no. 3, pp. 383-403. doi: 10.14736/kyb-2021-3-0383

Cité par Sources :