NeurIPS 2019 : Learning to Move - Walk Around

1 Travel Grants
1 Authorship/Co-Authorship

Welcome to the Learn to Move: Walk Around challenge, one of the official challenges in the NeurIPS 2019 Competition Track. Your task is to develop a controller for a physiologically plausible 3D human model to walk or run following velocity commands with minimum effort. You are provided with a human musculoskeletal model, a physics-based simulation environment, OpenSim, where you can synthesize physically and physiologically plausible motions, and additional tools (e.g. human gait data, lighter/faster simulation environment, and example codes) that will be useful for various learning approaches. There will be three tracks: 1) “the highest reward”, 2) “Novel ML Solution”, and 3) “Novel Biomechanical Solution.” You can compete for all of the tracks, where you will need to submit a paper describing the novelty of your solution for the latter two tracks.

This page will be updated with more details in early June. Meanwhile, you can explore the simulation environment (github repo) of our past competitions NeurIPS 2018: AI for Prosthetics Challenge and NIPS 2017: Learning to Run, which will be a basis for this year’s environment.

Learn to Move

Rough Timeline

  • Competition details: early June
  • Additional tools and documentation: late June
  • Competition open: late June
  • Round 1: ~ late October
  • Round 2: ~ early November
  • Winners announcement: ~ late November

Contact Us


TechCrunch Stanford News IEEE