{"ID":2833367,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.03538","arxiv_id":"2512.03538","title":"AdaPower: Specializing World Foundation Models for Predictive Manipulation","abstract":"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.","short_abstract":"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 computati...","url_abs":"https://arxiv.org/abs/2512.03538","url_pdf":"https://arxiv.org/pdf/2512.03538v1","authors":"[\"Yuhang Huang\",\"Shilong Zou\",\"Jiazhao Zhang\",\"Xinwang Liu\",\"Ruizhen Hu\",\"Kai Xu\"]","published":"2025-12-03T07:59:00Z","proceeding":"cs.RO","tasks":"[\"cs.RO\"]","methods":"[]","has_code":false}