1. Maximizing multi-information
- Creator:
- Ay, Nihat and Knauf, Andreas
- Format:
- bez média and svazek
- Type:
- model:article and TEXT
- Subject:
- multi-information, exponential family, relative entropy, pair-interaction, infomax principle, Boltzmann machine, and neural networks
- Language:
- English
- Description:
- Stochastic interdependence of a probability distribution on a product space is measured by its Kullback-Leibler distance from the exponential family of product distributions (called multi-information). Here we investigate low-dimensional exponential families that contain the maximizers of stochastic interdependence in their closure. Based on a detailed description of the structure of probability distributions with globally maximal multi-information we obtain our main result: The exponential family of pure pair-interactions contains all global maximizers of the multi-information in its closure.
- Rights:
- http://creativecommons.org/publicdomain/mark/1.0/ and policy:public