6 papers
MOSS: Mode-Oriented Source Shaping for Robot Flow Matching Policies
Zirun Ma, Jindou Jia, Tuo An +3
In robotic manipulation, the same observation often admits multiple feasible actions. Generative policies such as diffusion and flow matching naturally capture such multimodal beha…
Learning Skills from Historical Action Trajectories: Action Experience Dictionary for World Action Models
Qi Lyu, Jiahua Dong, Hao Shen +9
World Action Models (WAMs) couple visual dynamics prediction with action generation, yet they do not explicitly support the reuse of action experience across manipulation tasks. Fu…
LEEVLA: Seeing What Matters in Latent Environment Evolution for Vision-Language-Action
Qi Lyu, Baicheng Liu, Xudong Wang +3
Vision-language-action (VLA) models aim to map multimodal inputs to robot actions. However, most existing approaches struggle to cover complex dynamic scenarios due to treating all…
Lifelong Embodied Navigation Learning
Xudong Wang, Jiahua Dong, Baichen Liu +3
Embodied navigation agents powered by large language models have shown strong performance on individual tasks but struggle to continually acquire new navigation skills, which suffe…
Lifelong Language-Conditioned Robotic Manipulation Learning
Xudong Wang, Zebin Han, Zhiyu Liu +5
Traditional language-conditioned manipulation agent sequential adaptation to new manipulation skills leads to catastrophic forgetting of old skills, limiting dynamic scene practica…
Rehearsal-free Federated Domain-incremental Learning
Rui Sun, Haoran Duan, Jiahua Dong +3
We introduce a rehearsal-free federated domain incremental learning framework, RefFiL, based on a global prompt-sharing paradigm to alleviate catastrophic forgetting challenges in…