A compositional diffusion transformer generates synthetic robotic transitions for unseen task combinations and refines itself iteratively via offline RL validation, achieving high success on held-out tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Iterative Compositional Data Generation for Robot Control
A compositional diffusion transformer generates synthetic robotic transitions for unseen task combinations and refines itself iteratively via offline RL validation, achieving high success on held-out tasks.