Minimax entropy: The statistical physics of optimal models
David P. Carcamo, Nicholas J. Weaver, Purushottam D. Dixit, Christopher W. Lynn · 2025
Read Minimax entropy: The statistical physics of optimal models by David P. Carcamo, Nicholas J. Weaver, Purushottam D. Dixit, Christopher W. Lynn (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-2505-01607.txt.
Description
David Carcamo and co-authors ask a basic question about building models: which details from the data should a model keep? A good model leaves out as much as it can while staying accurate. Picking the right features to keep, though, has mostly been done by feel.
Using a principle that favors the shortest full account of the data, the team shows the best features are a special kind. They give the model that spreads its outcomes most widely while itself staying as tight as possible, a balance the authors call minimax entropy. The paper reviews uses from machine learning to models of living networks. It also notes that simple ways of finding such features fail on big data, so new math and computing ideas are needed.
Catalog record
| CREATOR | David P. Carcamo, Nicholas J. Weaver, Purushottam D. Dixit, Christopher W. Lynn |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-01607