Discrete Flow Maps
Peter Potaptchik, Jason Yim, Adhi Saravanan, Peter Holderrieth, Eric Vanden-Eijnden, Michael S. Albergo · 2026
Read Discrete Flow Maps by Peter Potaptchik, Jason Yim, Adhi Saravanan, Peter Holderrieth, Eric Vanden-Eijnden, Michael S. Albergo (2026) 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-2604-09784.txt.
Description
Potaptchik and five co-authors match side-by-side making with discrete data through Discrete Flow Maps. Stepwise guessing limits language models, while smooth flow models offer side-by-side making but demand looped math. Flow Maps compress paths into single mappings, making text from noise in one pass, but standard versions rely on flat losses ill-suited to discrete data. Discrete Flow Maps match path packing with chance-simplex shape, redoing training for the discrete domain. This shape fit lets the method surpass previous results in discrete flow models.
Catalog record
| CREATOR | Peter Potaptchik, Jason Yim, Adhi Saravanan, Peter Holderrieth, Eric Vanden-Eijnden, Michael S. Albergo |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77587 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.09784