Conservative Flows: A New Paradigm of Generative Models
Eshed Gal, Md Shahriar Rahim Siddiqui, Moshe Eliasof, Eldad Haber · 2026
Read Conservative Flows: A New Paradigm of Generative Models by Eshed Gal, Md Shahriar Rahim Siddiqui, Moshe Eliasof, Eldad Haber (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-2605-06905.txt.
Description
Eshed Gal, Md Shahriar Rahim Siddiqui, Moshe Eliasof, and Eldad Haber propose Conservative Flows, a generative modeling idea that starts from data instead of noise. Standard models transport random noise into samples through a learned flow. Conservative Flows instead run a step-by-step random dynamics that leaves the data distribution unchanged, seeded from data states, and can wrap around any trained flow model as its engine.
Two probability-preserving samplers are built: a corrected Langevin scheme with a Metropolis accept-reject step, and a predictor-corrector flow. Both run directly on existing checkpoints. On a spiral test target, ImageNet, and a flower dataset, the samplers steadily improve over the original generation methods they build on.
Catalog record
| CREATOR | Eshed Gal, Md Shahriar Rahim Siddiqui, Moshe Eliasof, Eldad Haber |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 55 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.06905