GORDON learns dense RL rewards from unlabeled video by embedding object-centric scene graphs, and uses the reward's temporal profile to automatically split long-horizon manipulation tasks into subtasks.
Maniskill3: Gpu parallelized robotics simulation and rendering for generalizable embodied ai,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
GORDON: Graph-based Object-centric Rewards for Decomposition of Long-Horizon Manipulation
GORDON learns dense RL rewards from unlabeled video by embedding object-centric scene graphs, and uses the reward's temporal profile to automatically split long-horizon manipulation tasks into subtasks.