A framework called Policy-as-Data generates task-oriented synthetic HOI data via RL policies in physics simulators, retargets it, and trains diffusion models that generalize to unseen objects and long horizons.
Zero-shot human-object interaction synthesis with multimodal priors
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
years
2026 4verdicts
UNVERDICTED 4representative citing papers
DeVI enables zero-shot physically plausible dexterous control by imitating synthetic videos via a hybrid 3D-human plus 2D-object tracking reward.
RePHO refines noisy kinematic HOI estimates from monocular videos into physically plausible sequences via RL in a physics simulator with adaptive dual self-updating sampling, showing metric gains on two benchmarks.
GRAIL creates over 20,000 synthetic loco-manipulation sequences from known 3D configurations and video priors, then trains policies that achieve 84% pick-up and 90% stair-climbing success on a real Unitree G1 humanoid using only the generated data.
citing papers explorer
-
Policy-as-Data: Learning Generalizable HOI Diffusion Models from Simulated Physics
A framework called Policy-as-Data generates task-oriented synthetic HOI data via RL policies in physics simulators, retargets it, and trains diffusion models that generalize to unseen objects and long horizons.
-
DeVI: Physics-based Dexterous Human-Object Interaction via Synthetic Video Imitation
DeVI enables zero-shot physically plausible dexterous control by imitating synthetic videos via a hybrid 3D-human plus 2D-object tracking reward.
-
Recovering Physically Plausible Human-Object Interactions from Monocular Videos
RePHO refines noisy kinematic HOI estimates from monocular videos into physically plausible sequences via RL in a physics simulator with adaptive dual self-updating sampling, showing metric gains on two benchmarks.
-
GRAIL: Generating Humanoid Loco-Manipulation from 3D Assets and Video Priors
GRAIL creates over 20,000 synthetic loco-manipulation sequences from known 3D configurations and video priors, then trains policies that achieve 84% pick-up and 90% stair-climbing success on a real Unitree G1 humanoid using only the generated data.