Research map / Video / Video generation
Do You Have Freestyle? Expressive Humanoid Locomotion via Audio Control
Zhe Li, Cheng Chi, Yangyang Wei, Boan Zhu, Tao Huang, Zhenguo Sun, Yibo Peng, Pengwei Wang, Zhongyuan Wang, Fangzhou Liu, Chang Xu, Shanghang Zhang
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 |
|---|---|---|---|
| use Delta-MoE teacher instead of vanilla MoE (ablation) | RoboPerform · BEAT2 IsaacGym | E_MPJPE: 0.14 -> 0.05 | Improved |
| use Delta-MoE teacher instead of vanilla MoE (ablation) | RoboPerform · FineDance MuJoCo | success rate: 0.61 -> 0.67 | Improved |
| add content latent from pretrained motion generator (ablation) | RoboPerform · BEAT2 IsaacGym | E_MPJPE: 0.11 -> 0.05 | Improved |
| add content latent (ablation) | RoboPerform · FineDance | E_MPJPE: 0.20 -> 0.18 | Mixed |
| add audio adaptor with InfoNCE alignment (ablation) | RoboPerform · BEAT2 IsaacGym | success rate: 0.88 -> 0.99 | Improved |
| add audio adaptor (ablation) | RoboPerform · FineDance MuJoCo | success rate: 0.51 -> 0.67 | Improved |
| add audio adaptor (ablation) | RoboPerform · FineDance IsaacGym | E_MPJPE: 0.49 -> 0.18 | Improved |
| retargeting-free direct audio policy vs generate-retarget-track baseline | RoboPerform · FineDance IsaacGym | success rate: 0.88 -> 0.93 | Improved |
| retargeting-free direct audio policy vs baseline | RoboPerform · BEAT2 IsaacGym | success rate: 0.98 -> 0.99 | No clear effect |
Closest papers by meaning
- RoboGesture: Real-Time Semantic-aligned Co-Speech Gestures Generation for Humanoid Interaction (ECCV 2026 · similarity 0.90)
- She Performs Your Voice: A Unified Speech and Dance Motion Model (NeurIPS 2026 · similarity 0.88)
- Audio-Driven Stylized Gesture Generation with Flow-Based Model (ECCV 2022 · similarity 0.88)
- From Language to Locomotion: Retargeting-free Humanoid Control via Motion Latent Guidance (ICLR 2026 · similarity 0.87)
- DiscoForcing: A Unified Framework for Real-Time Audio-Driven Character Control with Diffusion Forcing (ICML 2026 · similarity 0.86)
- 🎧MOSPA: Human Motion Generation Driven by Spatial Audio (NeurIPS 2025 · similarity 0.86)
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.
