Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2512.03538 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909941778874368 |
|---|---|
| author | Huang, Yuhang Zou, Shilong Zhang, Jiazhao Liu, Xinwang Hu, Ruizhen Xu, Kai |
| author_facet | Huang, Yuhang Zou, Shilong Zhang, Jiazhao Liu, Xinwang Hu, Ruizhen Xu, Kai |
| contents | World Foundation Models (WFMs) offer remarkable visual dynamics simulation capabilities, yet their application to precise robotic control remains limited by the gap between generative realism and control-oriented precision. While existing approaches use WFMs as synthetic data generators, they suffer from high computational costs and underutilization of pre-trained VLA policies. We introduce \textbf{AdaPower} (\textbf{Ada}pt and Em\textbf{power}), a lightweight adaptation framework that transforms general-purpose WFMs into specialist world models through two novel components: Temporal-Spatial Test-Time Training (TS-TTT) for inference-time adaptation and Memory Persistence (MP) for long-horizon consistency. Integrated within a Model Predictive Control framework, our adapted world model empowers pre-trained VLAs, achieving over 41\% improvement in task success rates on LIBERO benchmarks without policy retraining, while preserving computational efficiency and generalist capabilities. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_03538 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AdaPower: Specializing World Foundation Models for Predictive Manipulation Huang, Yuhang Zou, Shilong Zhang, Jiazhao Liu, Xinwang Hu, Ruizhen Xu, Kai Robotics World Foundation Models (WFMs) offer remarkable visual dynamics simulation capabilities, yet their application to precise robotic control remains limited by the gap between generative realism and control-oriented precision. While existing approaches use WFMs as synthetic data generators, they suffer from high computational costs and underutilization of pre-trained VLA policies. We introduce \textbf{AdaPower} (\textbf{Ada}pt and Em\textbf{power}), a lightweight adaptation framework that transforms general-purpose WFMs into specialist world models through two novel components: Temporal-Spatial Test-Time Training (TS-TTT) for inference-time adaptation and Memory Persistence (MP) for long-horizon consistency. Integrated within a Model Predictive Control framework, our adapted world model empowers pre-trained VLAs, achieving over 41\% improvement in task success rates on LIBERO benchmarks without policy retraining, while preserving computational efficiency and generalist capabilities. |
| title | AdaPower: Specializing World Foundation Models for Predictive Manipulation |
| topic | Robotics |
| url | https://arxiv.org/abs/2512.03538 |