AdaPower: Specializing World Foundation Models for Predictive Manipulation

TL;DR

AdaPower is a lightweight framework that adapts general world foundation models into specialized, efficient, and precise tools for robotic control, significantly improving task success rates without retraining policies.

Contribution

It introduces novel adaptation components, TS-TTT and MP, enabling effective specialization of WFMs for control tasks within an MPC framework.

Findings

Over 41% improvement in task success rates on LIBERO benchmarks

Achieves adaptation without retraining pre-trained policies

Maintains computational efficiency and generalist capabilities

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.