Research map / Reinforcement learning / Visual and representation RL
InterPrior: Scaling Generative Control for Physics-Based Human-Object Interactions
Sirui Xu, Samuel Schulter, Morteza Ziyadi, Xialin He, Xiaohan Fei, Yu-Xiong Wang, Liang-Yan Gui
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add RL finetuning (in-betweening to single-frame goals) to the distilled policy (ablation) | InterPrior · Multi-goal chaining (stress test) | success rate (%): 45.1 -> 68.8 | Improved |
| add RL finetuning to the distilled policy (ablation) | InterPrior · Random initialization lift (stress test) | success rate (%): 41.1 -> 88.6 | Improved |
| add RL finetuning to the distilled policy (ablation) | InterPrior · Snapshot goals | success rate (%): 89.1 -> 90.0 | Improved |
| add RL finetuning to the distilled policy (ablation) | InterPrior · Trajectory goals | success rate (%): 93.6 -> 94.6 | No clear effect |
| add latent shaping loss to InterMimic+ distilled policy (ablation) | InterPrior · Multi-goal chaining (stress test) | success rate (%): 33.9 -> 40.0 | Improved |
| add latent shaping loss (ablation) | InterPrior · Snapshot goals | success rate (%): 71.4 -> 74.9 | Improved |
| bound latent to hypersphere and bound observations (ablation) | InterPrior · Snapshot goals | success rate (%): 74.9 -> 89.1 | Improved |
| upgrade expert from InterMimic to InterMimic+ (perturbation, hand reward) (ablation) | MaskedMimic-style distillation · Contact goals | success rate (%): 52.2 -> 69.3 | Improved |
| use InterPrior instead of full-reference InterMimic for thin-object tracking | full-reference tracking · OMOMO thin objects + perturbation | success rate (%): 63.9 -> 83.2 SR; Eh 7.1 -> 8.9 | Mixed |
| finetune on new BEHAVE/HODome data starting from InterPrior rather than InterMimic | novel object adaptation · HODome | success rate (%): 55.5 -> 72.4 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- InterMimic — InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions (CVPR 2025) · better on 3 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- InterMimic: Towards Universal Whole-Body Control for Physics-Based Human-Object Interactions (CVPR 2025 · similarity 0.89)
- Universal Humanoid Motion Representations for Physics-Based Control (ICLR 2024 · similarity 0.89)
- Humanoid Horizon: Extending Task Horizon in Whole-Body Loco-Manipulation via Parallel Training, Dynamic Starting, and Reward Gating (NeurIPS 2026 · similarity 0.89)
- SUGAR: A Scalable Human-Video-Driven Generalizable Humanoid Loco-Manipulation Learning Framework (NeurIPS 2026 · similarity 0.88)
- Adversarial Locomotion and Motion Imitation for Humanoid Policy Learning (NeurIPS 2025 · similarity 0.88)
- Hierarchical World Models as Visual Whole-Body Humanoid Controllers (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
