Saved in:
Bibliographic Details
Main Authors: Carcamo, David P., Weaver, Nicholas J., Dixit, Purushottam D., Lynn, Christopher W.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.01607
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913816966594560
author Carcamo, David P.
Weaver, Nicholas J.
Dixit, Purushottam D.
Lynn, Christopher W.
author_facet Carcamo, David P.
Weaver, Nicholas J.
Dixit, Purushottam D.
Lynn, Christopher W.
contents When constructing models of the world, we aim for optimal compressions: models that include as few details as possible while remaining as accurate as possible. But which details -- or features measured in data -- should we choose to include in a model? Here, using the minimum description length principle, we show that the optimal features are the ones that produce the maximum entropy model with minimum entropy, thus yielding a minimax entropy principle. We review applications, which range from machine learning to optimal models of biological networks. Naive implementations, however, are limited to systems with small numbers of states and features. We therefore require new theoretical insights and computational techniques to construct optimal compressions of high-dimensional datasets arising in large-scale experiments.
format Preprint
id arxiv_https___arxiv_org_abs_2505_01607
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Minimax entropy: The statistical physics of optimal models
Carcamo, David P.
Weaver, Nicholas J.
Dixit, Purushottam D.
Lynn, Christopher W.
Quantitative Methods
When constructing models of the world, we aim for optimal compressions: models that include as few details as possible while remaining as accurate as possible. But which details -- or features measured in data -- should we choose to include in a model? Here, using the minimum description length principle, we show that the optimal features are the ones that produce the maximum entropy model with minimum entropy, thus yielding a minimax entropy principle. We review applications, which range from machine learning to optimal models of biological networks. Naive implementations, however, are limited to systems with small numbers of states and features. We therefore require new theoretical insights and computational techniques to construct optimal compressions of high-dimensional datasets arising in large-scale experiments.
title Minimax entropy: The statistical physics of optimal models
topic Quantitative Methods
url https://arxiv.org/abs/2505.01607