Modeling and Understanding Neural Weights of Humanoid Policies with Diffusion

Key points

• However, humanoid control networks are typically obtained through conventional training procedures, while comparatively little attention has been given to directly modeling the distribution of their learned weights and generating functional networks in weight space.1
• In this work, we investigate whether high-performing humanoid control networks can be generatively modeled directly through their weights.1
• We adopt a latent diffusion framework that couples a linear autoencoder with a Diffusion Transformer.1
• Trained on networks collected from reinforcement learning for the Unitree G1 robot, our approach generates deployable networks that perform various motions in the real world, including walking, dancing, and spin kick.1

Sources (1)

• [1]Modeling and Understanding Neural Weights of Humanoid Policies with Diffusion
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 05:59 PM

  Recent advances in robot learning have driven rapid progress in humanoid control, enabling increasingly agile behaviors in the real world.

  However, humanoid control networks are typically obtained through conventional training procedures, while comparatively little attention has been given to directly modeling the distribution of their learned weights and generating functional networks in weight space.

Extractive summary: sentences quoted from the sources.