Dream2Reward: Transition-Alignment Reward Models from Positive Demonstrations for Robotic Manipulation
arXiv:2608.18787v1 Announce Type: new Abstract: Learning robotic policies requires dense rewards that remain informative when behavior departs from successful demonstrations. Progress-based rewards estimate how far an observation has advanced along a nominal successful trajectory, but may remain high after an incorrect transition. We introduce Dream2Reward, which learns a language-conditioned successful latent transition field from positive demonstrations. Given the visual history up to a trans
Overview
arXiv:2608.18787v1 Announce Type: new Abstract: Learning robotic policies requires dense rewards that remain informative when behavior departs from successful demonstrations. Progress-based rewards estimate how far an observation has advanced along a nominal successful trajectory, but may remain high after an incorrect transition. We introduce Dream2Reward, which learns a language-conditioned successful latent transition field from positive demonstrations. Given the visual history up to a transition start, the model predicts the latent displacement associated with successful execution and scores the observed displacement through signed directional and symmetric magnitude agreement. This transition-level comparison penalizes wrong-direction, overshooting, and stagnant motion even when the resulting observation appears to show progress. Dream2Reward requires no failure annotations, progress labels, or synthetic negatives, and produces a dense causal reward. Across mechanism diagnostics and shared-trajectory evaluations, it provides stronger success-failure separation and more informative feedback on low-quality behavior than progress-based alternatives. Across online and offline policy learning, the same frozen reward model reduces reward hacking and supports stronger downstream performance, including in real-robot manipulation. These results show that comparing realized motion with predicted successful change provides an effective way to convert positive demonstrations into dense rewards for robot learning.
Source
Originally published at arxiv.org.
Related Articles
- Scheduling and Routing with Degradation-Triggered Job Arrivals: An Application to Forest Firefighting with an Unmanned Aerial Vehicle Fleet
- GAPL: Grounded Action-effect Policy Learning for LLM-Based Trajectory Planning
- Transferable Tool-Tissue Contact Detection from Stereo Depth in Robot-Assisted Surgery
Source: https://arxiv.org/abs/2608.18787