Summary of Wandr: Intention-guided Human Motion Generation, by Markos Diomataris et al.
WANDR: Intention-guided Human Motion Generation
by Markos Diomataris, Nikos Athanasiou, Omid Taheri, Xi Wang, Otmar Hilliges, Michael J. Black
First submitted to arxiv on: 23 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
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 This paper addresses the long-standing challenge of synthesizing natural human motions that enable a 3D human avatar to walk and reach for arbitrary goals in 3D space. Current methods, including data-driven and reinforcement learning approaches, are limited by their inability to generalize and produce naturally looking movements. The primary obstacle is the scarcity of training data that combines locomotion with goal reaching. To overcome this challenge, the authors introduce WANDR, a data-driven model that generates natural human motions based on an avatar’s initial pose and a goal’s 3D position. This conditional Variational Auto-Encoder (c-VAE) incorporates novel intention features that drive rich goal-oriented movement, allowing it to adapt to novel situations without requiring sub-goals or entire motion paths. The authors evaluate WANDR extensively using the AMASS and CIRCLE datasets, demonstrating its ability to generate natural and long-term motions that reach 3D goals and generalize to unseen locations. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us create more realistic and useful computer-controlled humans in virtual worlds. Right now, it’s hard to make these avatars move naturally and reach for things they want. The authors are trying to fix this problem by creating a special computer program that can generate human-like movements based on where the avatar wants to go. They’re using a new way of teaching machines called intention features, which helps them learn how to move in a more goal-oriented way. This means the avatars can adapt to new situations and reach for things without needing specific instructions. The authors tested their program with different types of data and showed that it can make realistic movements that work even when the goal is far away or hard to reach. |
Keywords
» Artificial intelligence » Encoder » Reinforcement learning
