From Digital Human Interactions to Physics-Based Humanoid Skills: Physics-Grounded Post-Training of Interaction Generators

Key points

• Recent methods have made promising progress in generating interactions between two humanoids, largely relying on physics-based tracking policies to convert digital reference motions into executable trajectories.1
• Our DIGHT first executes multiple text-conditioned interaction candidates in simulation using a fixed tracker.1
• Rather than collapsing these signals into a single scalar reward for candidate ranking, we align the pretrained generator using physics-decoupled diffusion direct preference optimization (DPO), preserving criterion-specific supervision without differentiating through the simulator.1
• Additionally, to improve interaction fidelity, we propose to incorporate force feedback from simulator as a measure of contact fidelity and construct preferences over contact occurrence, location, duration, and force magnitude.1

Sources (1)

• [1]From Digital Human Interactions to Physics-Based Humanoid Skills: Physics-Grounded Post-Training of Interaction Generators
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 7, 04:12 PM

  In this paper, we introduce DIGHT, a co-adaptive framework that couples a Digital human Interaction Generator with a Humanoid Tracking policy.

  Recent methods have made promising progress in generating interactions between two humanoids, largely relying on physics-based tracking policies to convert digital reference motions into executable trajectories.

Extractive summary: sentences quoted from the sources.