TY  - RPRT
TI  - SOLE-R1: Video-Language Reasoning as the Sole Reward for On-Robot Reinforcement Learning
AU  - Philip Schroeder
AU  - Thomas Weng
AU  - Karl Schmeckpeper
AU  - Eric Rosen
AU  - Stephen Hart
AU  - Ondrej Biza
PY  - 2026
UR  - https://arxiv.org/abs/2603.28730
ID  - 2603.28730
ER  -