DREAMSTEER: Latent World Models Enable VLA Deployment Robustness Without FinetuningDREAMSTEER：潜在世界模型支持VLA无微调部署鲁棒性
S 1.7T11 sources1 个来源R7-research
DREAMSTEER is a deployment-time steering framework for pretrained vision-language-action policies that achieves robustness without requiring finetuning or parameter modifications.
The system leverages latent world models and language-conditioned value models to evaluate candidate actions, addressing the distribution shift problem between training and deployment environments.
Tested on four real-world manipulation benchmarks with unseen objects, DREAMSTEER demonstrates significant improvements in task success rates.