DynamicTHOR: A Scalable Dataset of Human-Centric Dynamic Scenes for Embodied AI
Chenxu Wang, Dunzheng Wang, Dong Wang, Xinghang Li, Yongchang Li, Huaping Liu · 2026
Read DynamicTHOR: A Scalable Dataset of Human-Centric Dynamic Scenes for Embodied AI by Chenxu Wang, Dunzheng Wang, Dong Wang, Xinghang Li, Yongchang Li, Huaping Liu (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/pmc-13365494.txt.
Description
Chenxu Wang and co-authors build a dataset of homes where people move things around. Their 2026 paper answers a gap in embodied AI: robots train in still rooms, but real homes shift as residents live in them. Hand-building moving scenes is slow and dear.
Language models simulate residents and their object shuffling inside virtual homes. The result holds fifty characters and one hundred dynamic scenes, with room to grow. A user study rated the generated motion as believable as human-made versions. Navigation in changing scenes is the first task in view. The authors release the framework so others can scale it further.
Catalog record
| CREATOR | Chenxu Wang, Dunzheng Wang, Dong Wang, Xinghang Li, Yongchang Li, Huaping Liu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |