Compositional Humanoid Manipulation through Isomorphic Teleoperation and Independent Skill Routing
We present Humanoid Semantic Mixture-of-Experts Diffusion Policy (HS-MoDP), which composes separately demonstrated arm skills through independent semantic routing over a shared expert pool.
ProofPaper ↗
Key points
- Humanoid manipulation requires reusable arm skills, yet demonstrating every possible skill combination is impractical.
- A semantic scheduler integrates the learned policy with an existing locomotion controller through alternating manipulation and repositioning stages, extending the effective workspace and enabling skill reuse across location changes.
- To acquire demonstrations of whole-body contact manipulation, we develop G1Exo, an isomorphic exoskeleton that provides direct joint-level control of arm configurations.
- G1Exo enables more reliable contact teleoperation than end-effector VR-IK, and its demonstrations improve contact-skill learning across four policy architectures.
Sources (1)
- [1]Compositional Humanoid Manipulation through Isomorphic Teleoperation and Independent Skill RoutingarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 10:59 AM
We present Humanoid Semantic Mixture-of-Experts Diffusion Policy (HS-MoDP), which composes separately demonstrated arm skills through independent semantic routing over a shared expert pool.
Humanoid manipulation requires reusable arm skills, yet demonstrating every possible skill combination is impractical.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 19, 2026TorchNeuron: running HF models on AWS Trainium is easier than ever
- Oct 18, 2026Comfy-Org/ComfyUI v0.42.0
- Oct 17, 2026Nemotron 3.5 Bolt, Muse Spark Open, and a Fast LFM2.5 24B
- Oct 16, 2026[Research] 2,788 tokens/s on a single RTX 5090 with a diffusion LLM — Introducing ORL (fully open source)
- Oct 16, 2026Grounding AI agents in biological reality: Building blocks for autonomous wet labs
- Oct 16, 2026huggingface/transformers v5.20.1: Release v5.20.1