Low-Rank Adaptation (LoRA) is proposed as a structural regularizer for critic learning from scratch that reduces critic loss during training and improves overall policy performance, achieving the best or competitive results on most tasks.
Yuan Zhuang, Yuexin Bian, Si-Hong He et al.· arXiv.org· 0 citations
Prior-Informed Odometry from Human-Motion Tracking (PRIMO) generates odometry supervision by having the humanoid track diverse retargeted human motions in simulation, decoupling supervision from the deployment policies and broadening the training motion distribution, and consistently lowers mean trajectory errors relat...
Xuan-Wen Han, Ang Li, Shaoxiong Zhang et al.· 0 citations
Artificial Foveated Perception is proposed, a lightweight, policy-agnostic module that takes the same vision and language inputs as Vision-Language-Action and World Action Model pipelines and predicts task-conditioned masks over relevant objects, the robot, and other action-critical regions and reduces fine-tuning time...