Showing cs.ROShow all
2 papers · 1 filter
cs.RO2026
ReAT: Residual Action Tokenization for Autoregressive Robot Policies
Dijia Zhan, Zhuo Li, Xuemiao Xu +4
Autoregressive robot policies rely on action tokenizers to convert continuous action chunks into compact sequences of discrete prediction targets. Existing tokenizers can reconstru…
cs.RO2026
SEAM: Smooth Execution of Action-Chunked Motion for Vision-Language-Action Policies
Dijia Zhan, Xuemiao Xu, Jinyi Li +1
Vision-Language-Action (VLA) policies that execute fixed-length action chunks can exhibit multimodal bifurcation: a cross-chunk inconsistency in which adjacent chunks generated fro…