OGPO is a sample-efficient off-policy method for full finetuning of generative control policies that reaches SOTA on robotic manipulation tasks and can recover from poor behavior-cloning initializations without expert data.
Flow policy gradients for robot control
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 3verdicts
UNVERDICTED 3roles
background 1polarities
background 1representative citing papers
PODPO is a likelihood-free generative policy optimization method for online RL that steers actions to high-return regions using only positive-advantage samples and local contrastive drifting.
ConSFT is a gradient-scaling fine-tuning objective for flow-matching VLAs that bounds parameter disruption via model-confidence weighting, yielding over 20% better capability retention than vanilla SFT on LIBERO and RoboTwin.
citing papers explorer
-
OGPO: Sample Efficient Full-Finetuning of Generative Control Policies
OGPO is a sample-efficient off-policy method for full finetuning of generative control policies that reaches SOTA on robotic manipulation tasks and can recover from poor behavior-cloning initializations without expert data.
-
Positive-Only Drifting Policy Optimization
PODPO is a likelihood-free generative policy optimization method for online RL that steers actions to high-return regions using only positive-advantage samples and local contrastive drifting.
-
Preserving Foundational Capabilities in Flow-Matching VLAs through Conservative SFT
ConSFT is a gradient-scaling fine-tuning objective for flow-matching VLAs that bounds parameter disruption via model-confidence weighting, yielding over 20% better capability retention than vanilla SFT on LIBERO and RoboTwin.