POGP trains a prefix value function over the diffusion denoising chain, giving a learned early-stopping rule that cuts average denoising steps about 2.7x with 98-99% of full-chain return and a small gain in final task performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning When to Stop: Prefix-Optimal Dynamic Diffusion Policies for Continuous Control
POGP trains a prefix value function over the diffusion denoising chain, giving a learned early-stopping rule that cuts average denoising steps about 2.7x with 98-99% of full-chain return and a small gain in final task performance.