Map-World: Masked Action planning and Path-Integral World Model for Autonomous Driving
Bin Hu, Zijian Lu, Haicheng Liao, Chengran Yuan, Bin Rao, Yongkang Li, Guofa Li, Zhiyong Cui, Cheng-zhong Xu, Zhenning Li · 2025
Read Map-World: Masked Action planning and Path-Integral World Model for Autonomous Driving by Bin Hu, Zijian Lu, Haicheng Liao, Chengran Yuan, Bin Rao, Yongkang Li, Guofa Li, Zhiyong Cui, Cheng-zhong Xu, Zhenning Li (2025) 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-2511-20156.txt.
Description
Hu, Lu, Liao, and colleagues plan self-driving motion without hand-made trajectory anchors. Their 2025 paper presents MAP-World for autonomous driving. Planners must weigh several plausible futures at once. Anchor libraries and reinforcement selectors discard alternatives early. That narrowing complicates training and loses information.
Masked Action Planning treats future motion as sequence completion. Past waypoints stay visible as ordinary tokens. Future waypoints begin as mask tokens to fill. A driving-intent path scaffolds the completion coarsely. Noise injection spawns multiple diverse trajectory candidates. A compact world model rolls out future road semantics. Each candidate trajectory conditions its own rollout. Training averages semantic losses across all modes. Trajectory probabilities weight that expectation like path integrals. On NAVSIM the method matched anchor-based rivals. Among world-model planners it reached state-of-the-art results. Inference stays fast enough for real-time use. Fixed mode counts remain a noted limitation.
Catalog record
| CREATOR | Bin Hu, Zijian Lu, Haicheng Liao, Chengran Yuan, Bin Rao, Yongkang Li, Guofa Li, Zhiyong Cui, Cheng-zhong Xu, Zhenning Li |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-20156