aiqbench analysis
BayesFP introduces a retraining-free method that applies Feynman-Kac sampling to constrain diffusion and flow-matching policies for robot trajectory generation without sacrificing faithfulness to expert demonstrations. This approach enables zero-shot adaptation to new safety constraints and non-convex obstacles, offering a practical solution for deploying pretrained generative models in dynamic robotic environments.
BayesFP: Posterior Estimation for Flow-Based Policies via Feynman-Kac Sampling
arXiv:2606.21014v2 Announce Type: replace Abstract: Robots must generate trajectories that remain faithful to learned expert behavior while satisfying safety constraints and task-specific objectives specified only at inference time. We formulate constrained trajectory generation for pretrained diffusion and flow-matching policies as Bayesian posterior sampling, with the learned demonstration distribution as a prior and an inference-time, cost-derived likelihood tilting it toward feasible, optimal trajectories. To sample from this posterior without any retraining of the base policy, we leverage the Feynman--Kac corrector framework, originally formulated for diffusion models, and extend it to deterministic flow-matching policies. The result is a unified, inference-time, retraining-free sampler for diffusion and flow policies. We validate the approach on pretrained Diffusion Policy, GR00T-N1.6, and $\pi_{0.5}$ checkpoints across simulated and real-world manipulation tasks, including planning around non-convex obstacles introduced at inference time, and show improvements over the base $\pi_{0.5}$ on zero-shot tasks.
Abstract:Robots must generate trajectories that remain faithful to learned expert behavior while satisfying safety constraints and task-specific objectives specified only at inference time. We formulate constrained trajectory generation for pretrained diffusion and flow-matching policies as Bayesian posterior sampling, with the learned demonstration distribution as a prior and an inference-time, cost-derived likelihood tilting it toward feasible, optimal trajectories. To sample from this posterior without any retraining of the base policy, we leverage the Feynman--Kac corrector framework, originally formulated for diffusion models, and extend it to deterministic flow-matching policies. The result is a unified, inference-time, retraining-free sampler for diffusion and flow policies. We validate the approach on pretrained Diffusion Policy, GR00T-N1.6, and $\pi_{0.5}$ checkpoints across simulated and real-world manipulation tasks, including planning around non-convex obstacles introduced at inference time, and show improvements over the base $\pi_{0.5}$ on zero-shot tasks.
Robotics (cs.RO)
arXiv:2606.21014 [cs.RO]
(or arXiv:2606.21014v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.21014 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 228 words
Related coverage
GRPO-QM: Target Preserving Exploration for Quantum Tomography
arXiv · Sep 15, 2026
Amortized Score-Hamiltonian Policy Iteration: A Grid-Free Scheme for Relaxed Stochastic Control Problems
arXiv · Oct 6, 2026
Conditional Mean Flows: Few-Step Densities for Simulation-Based Inference
arXiv · Oct 16, 2026
[AINews] not much happened today
Latent Space · Sep 18, 2026