Conservative Offline Robot Policy Learning via Posterior-Transition Reweighting

TL;DR

This paper introduces Posterior-Transition Reweighting (PTR), a novel offline robot policy adaptation method that selectively emphasizes more reliable training samples, improving robustness when dealing with heterogeneous datasets.

Contribution

The paper proposes PTR, a reward-free, conservative post-training approach that reweights training samples based on their attributable post-action consequences, enhancing offline robot policy learning.

Findings

PTR improves policy robustness on heterogeneous datasets.

The method is compatible with diffusion and flow-matching action heads.

PTR outperforms uniform weighting in offline adaptation tasks.

Abstract

Offline post-training adapts a pretrained robot policy to a target dataset by supervised regression on recorded actions. In practice, robot datasets are heterogeneous: they mix embodiments, camera setups, and demonstrations of varying quality, so many trajectories reflect recovery behavior, inconsistent operator skill, or weakly informative supervision. Uniform post-training gives equal credit to all samples and can therefore average over conflicting or low-attribution data. We propose Posterior-Transition Reweighting (PTR), a reward-free and conservative post-training method that decides how much each training sample should influence the supervised update. For each sample, PTR encodes the observed post-action consequence as a latent target, inserts it into a candidate pool of mismatched targets, and uses a separate transition scorer to estimate a softmax identification posterior over target indices. The posterior-to-uniform ratio defines the PTR score, which is converted into a clipped-and-mixed weight and applied to the original action objective through self-normalized weighted regression. This construction requires no tractable policy likelihood and is compatible with both diffusion and flow-matching action heads. Rather than uniformly trusting all recorded supervision, PTR reallocates credit according to how attributable each sample's post-action consequence is under the current representation, improving conservative offline adaptation to heterogeneous robot data.