DreamControl-v2: Simpler and Scalable Autonomous Humanoid Skills via Trainable Guided Diffusion Priors
Sudarshan Harithas, Sangkyung Kwak, Pushkal Katara, Srujan Deolasee, Dvij Kalaria, Srinath Sridhar, Sai Vemprala, Ashish Kapoor, Jonathan Chung-Kuan Huang · 2026
Read DreamControl-v2: Simpler and Scalable Autonomous Humanoid Skills via Trainable Guided Diffusion Priors by Sudarshan Harithas, Sangkyung Kwak, Pushkal Katara, Srujan Deolasee, Dvij Kalaria, Srinath Sridhar, Sai Vemprala, Ashish Kapoor, Jonathan Chung-Kuan Huang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-00202.txt.
Description
Harithas and eight colleagues improve DreamControl for humanoid loco-manipulation by training a guided diffusion model in the humanoid's motion space. Pooling varied human and robot data into one embodiment space captures a wider skill range and removes manual filter steps. Scaling up the making of reference trajectories matters for robust downstream reinforcement-learning policies. Tests in simulation and with a real Unitree-G1 confirm the approach.
Catalog record
| CREATOR | Sudarshan Harithas, Sangkyung Kwak, Pushkal Katara, Srujan Deolasee, Dvij Kalaria, Srinath Sridhar, Sai Vemprala, Ashish Kapoor, Jonathan Chung-Kuan Huang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49110 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.00202