Flow Matching with General Discrete Paths: A Kinetic-Optimal Perspective
Neta Shaul, Itai Gat, Marton Havasi, Daniel Severo, Anuroop Sriram, Peter Holderrieth, Brian Karrer, Yaron Lipman, Ricky T. Q. Chen · 2024
Read Flow Matching with General Discrete Paths: A Kinetic-Optimal Perspective by Neta Shaul, Itai Gat, Marton Havasi, Daniel Severo, Anuroop Sriram, Peter Holderrieth, Brian Karrer, Yaron Lipman, Ricky T. Q. Chen (2024) 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-2412-03487.txt.
Description
Neta Shaul and co-authors widen the craft space for neat-data build tools, the kind that make text, shots, or new stuff from noise. Past neat builds lean on one plain spoil path: mask and fill. Their 2024 work lets any spoil path drive the build, through chains that hop state to state in smooth time. Speed laws are writ for each path through a shared lens: least waste of move force.
Path and speed split clean, so a user can set the spoil to fit the field. One blend path proves best by that force mark. Races span text, rock stuff, and shots. The blend path beats mask builds even in text, the mask's home turf. Free path picks can thus pay past one house style.
Catalog record
| CREATOR | Neta Shaul, Itai Gat, Marton Havasi, Daniel Severo, Anuroop Sriram, Peter Holderrieth, Brian Karrer, Yaron Lipman, Ricky T. Q. Chen |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 90 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-03487