Summary of Hierarchical World Models As Visual Whole-body Humanoid Controllers, by Nicklas Hansen et al.
Hierarchical World Models as Visual Whole-Body Humanoid Controllers
by Nicklas Hansen, Jyothir S V, Vlad Sobal, Yann LeCun, Xiaolong Wang, Hao Su
First submitted to arxiv on: 28 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The abstract presents a reinforcement learning approach for visual whole-body control of humanoids, leveraging hierarchical world models and rewards to train agents. The authors propose a framework that learns from visual observations without simplifying assumptions or skill primitives, achieving high-performance control policies in 8 tasks with a simulated 56-DoF humanoid. The approach synthesizes motions preferred by humans. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper shows how reinforcement learning can be used to control humanoids, which is challenging because they have many joints and are prone to falling over. To make this easier, the authors use a hierarchical model that combines high-level decision-making with low-level action execution. This approach is trained using rewards, allowing it to learn complex behaviors. The results show that the system can perform well in 8 different tasks, such as walking or reaching for objects. |
Keywords
* Artificial intelligence * Reinforcement learning
