# MetaWorld: Skill Transfer and Composition in a Hierarchical World Model for Grounding High-Level Instructions

arXiv: https://arxiv.org/abs/2601.17507 · HTML: https://mlsanity.com/paper/2601.17507/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Humanoid and legged locomotion world models, one of 11 directions in World models.
- Builds on 3 of its 23 references that are in the map.
- First posted on arXiv 24 Jan 2026.

## Placed in

- World models / [Humanoid and legged locomotion world models](https://mlsanity.com/wm/humanoid-and-legged-locomotion-world-models/)

## Builds on (most cited first, 3 of 3)

- [Training Agents Inside of Scalable World Models](https://mlsanity.com/paper/2509.24527/) (2025)
- [HumanoidBench: Simulated Humanoid Benchmark for Whole-Body Locomotion and Manipulation](https://mlsanity.com/paper/2403.10506/) (2024)
- [DETACH: Cross-domain Learning for Long-Horizon Tasks via Mixture of Disentangled Experts](https://mlsanity.com/paper/2508.07842/) (2025)

## Built on by (most cited first, 1 of 1)

- [From Digital Twins to World Models:Opportunities, Challenges, and Applications for Mobile Edge General Intelligence](https://mlsanity.com/paper/2603.17420/) (2026)

## Abstract (arXiv)

Humanoid robot loco-manipulation remains constrained by the semantic-physical gap. Current methods face three limitations: Low sample efficiency in reinforcement learning, poor generalization in imitation learning, and physical inconsistency in VLMs. We propose MetaWorld, a hierarchical world model that integrates semantic planning and physical control via expert policy transfer. The framework decouples tasks into a VLM-driven semantic layer and a latent dynamics model operating in a compact state space. Our dynamic expert selection and motion prior fusion mechanism leverages a pre-trained multi-expert policy library as transferable knowledge, enabling efficient online adaptation via a two-stage framework. VLMs serve as semantic interfaces, mapping instructions to executable skills and bypassing symbol grounding. Experiments on Humanoid-Bench show MetaWorld outperforms world model-based RL in task completion and motion coherence. Our code will be found at https://anonymous.4open.science/r/metaworld-2BF4/