Self-Refining Video Sampling
Sangwon Jang, Taekyung Ki, Jaehyeong Jo, Saining Xie, Jaehong Yoon, Sung Ju Hwang · 2026
Read Self-Refining Video Sampling by Sangwon Jang, Taekyung Ki, Jaehyeong Jo, Saining Xie, Jaehong Yoon, Sung Ju Hwang (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-2601-18577.txt.
Description
Sangwon Jang and five colleagues teach a video tool to check and fix its own work as it builds each clip. New video tools can make fine clips, but their moves often break the laws of how stuff moves. A ball may float, or a cup may pass through a desk. Past fixes call in a judge tool or train the tool on more clips. Both paths cost a lot and still miss fine slips.
Their plan lets the tool sample a clip, grade its own moves, and sample again with the grade in mind. The loop needs no new training and no judge from out-side. A self-check loop could lift the truth of moves in clips for film, ads, and games. It could also cut the cost of each good clip.
Catalog record
| CREATOR | Sangwon Jang, Taekyung Ki, Jaehyeong Jo, Saining Xie, Jaehong Yoon, Sung Ju Hwang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 77 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-18577