Future Representation Distillation for Imagination-Free World Action Models
Overview
This paper introduces DreamFree, a World Action Model (WAM) that repurposes a pretrained video diffusion model as a predictive visual encoder through representation distillation. By predicting the frozen model's representations from current observations alone, DreamFree enables action prediction without generating future videos. The method achieves real-time closed-loop control at 11Hz with a 14B video diffusion backbone and outperforms state-of-the-art VLAs and WAMs across real-world and simulated tasks. It also supports learning new robot tasks from video demonstrations of different embodiments without requiring action annotations for those new tasks.
Generated from attributed reports · Updated 1 hours ago
Event evidence and corrections
0 attributed source owners. Ownership does not establish independent confirmation. Quantities are reported separately and are never added together.
Reported quantity · control frequency: 11 other · Basis not reported
“relevant information from the visual encoder. Across real-world and simulated tasks, DreamFree achieves higher success rates than state-of-the-art VLAs and WAMs, while performing real-time closed-loop control at 11Hz with a 14B video diffusion backbone. Experiments show that DreamFree can translate stronger video priors and additional video data into better robot control and enable learning new robot tasks from video”
Exact source · revision 1Source owner not reported
Report timeline
Follow attributed reports and material updates.
- arXiv Robotics — research abstractsSignalFuture Representation Distillation for Imagination-Free World Action Models
This paper introduces DreamFree, a World Action Model (WAM) that repurposes a pretrained video diffusion model as a predictive visual encoder through representation distillation. By predicting the frozen model's representations from current observations alone, DreamFree enables action prediction without generating future videos. The method achieves real-time closed-loop control at 11Hz with a 14B video diffusion backbone and outperforms state-of-the-art VLAs and WAMs across real-world and simulated tasks. It also supports learning new robot tasks from video demonstrations of different embodiments without requiring action annotations for those new tasks.
Event coverage history
There is not enough continuous observation data to show a trend.
Timezone · UTC
Article dates follow your selected timezone. Briefing editions use Hong Kong time (UTC+8).