W2A: Distilling World Priors into Action Models for Real-Time Robotic Control
W2A: Distilling World Priors into Action Models for Real-Time Robotic Control
This research introduces W2A, a two-stage distillation framework designed to enable real-time robotic manipulation by transferring intermediate denoising representations from a video diffusion model into an action model during training. The approach discards the video model at test time to reduce computational overhead and latency. Experiments in both simulation and real-world settings show that W2A achieves a 95.9% average success rate on the LIBERO benchmark and approximately 21x faster inference compared to the representative World Action Model (WAM) baseline.
The paper proposes W2A, a two-stage distillation framework that transfers intermediate denoising representations from a video diffusion model into an action model to enable real-time robotic control without the video model at test time, achieving a 95.9% average success rate on LIBERO and approximately 21x faster inference than the WAM baseline in both simulation and real-world settings.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
inference speed improvement
≈21
View original evidence
fective supervision for action generation. This makes it possible to transfer such information effectively into an action model. Extensive experiments in both simulation and real-world settings demonstrate competitive success rates while significantly improving inference efficiency. On LIBERO, W2A achieves a 95.9% average success rate. More importantly, W2A achieves about 21x faster inference than the representative
Open source S4
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
fective supervision for action generation. This makes it possible to transfer such information effectively into an action model. Extensive experiments in both simulation and real-world settings demonstrate competitive success rates while significantly improving inference efficiency. On LIBERO, W2A achieves a 95.9% average success rate. More importantly, W2A achieves about 21x faster inference than the representative
Open source S4
Source：arXiv Robotics — research abstracts · arxiv.org