MyoChallenge'23
Embedded Files
Towards Human-Level Dexterity and Agility
Introducing MyoChallenge 2023: Towards Human-Level Dexterity and Agility
Introducing MyoChallenge 2023: Towards Human-Level Dexterity and Agility
Humans effortlessly grasp objects of diverse shapes and properties and execute agile locomotion without overwhelming their cognitive capacities. This ability was acquired through millions of years of evolution, which honed the symbiotic relationship between the central and peripheral nervous systems and the musculoskeletal structure. Consequently, it is not surprising that uncovering the intricacies of these complex, evolved systems underlying human movement remains a formidable challenge. Advancements in neuromechanical simulations and data driven methods offer promising avenues to overcome these obstacles. This year’s competition will feature two tracks: the manipulation track and the locomotion track.
Competition Tracks
Competition Tracks
The MyoChallenge consists of two tracks:
The MyoChallenge consists of two tracks:
Manipulation:
Develop a general manipulation policy capable of interacting with common household objects,such as children’s toys.
Locomotion:
Resembling the World Chase Tag competition, the MyoLeg musculoskeletal model is required to chase or evade an opponent in a 12 x 12-meter arena.
Competition Stages + Timeline
The entire competition is split into two stages. This helps us calibrate the difficulty of the challenge and avoid overfitting.
Competition Stages + Timeline
The entire competition is split into two stages. This helps us calibrate the difficulty of the challenge and avoid overfitting.
Phase 1: (July 13th - October 4th)
This phase focuses on problems requiring minimal generalization. The goal of this phase is to bootstrap teams on MyoChallenge problem statements. This phase serves as a qualifier for the next phase. Each team is constrained to submit a maximum of 5 solutions per day and automatically ranked on a leaderboard.
Phase 2: (October 5th - November 6th)
Teams that meet the qualifying criteria of phase-1 will be participating in phase-2. This phase will focus on task generalization and is the final battleground for the MyoChallenge'23 leading to the final leaderboard of the competition. Each participant will be constrained to submit a maximum 5 solutions per day and automatically ranked on a leaderboard.
Results: (December 12th)
Results announcement @ NeurIPS 23
Workshop @ NeurIPS 23: (December 16th)
How to start
How to start
MyoChallenge is built upon the MyoSuite ecosystem that consists of a set of physiologically realistic musculoskeletal models and a suite of tasks defined in OpenAI gym interface. The MyoSuite hosts a diverse collection of tasks, associated pre-trained behaviors policies, and training code to replicate results. If you're interested in participating, take a look at our challenge repo to get an overview on what to do next.
How to Participate
How to Participate
Awards
Awards
Awards Winners:
Awards Winners:
Winner Manipulation Track: Team Lattice
Winner Locomotion Track: Team GaitNet
Best Student Teams Award: Team JustRandom
DEI: Team NUABILITY
MyoChallenge'23 Leaderboard
MyoChallenge'23 Leaderboard
Manipulation:
Manipulation:
Phase 1:
Rank Participant team Score (↑) Effort (↓)
1 GaitNet 1.00000 0.02720
2 NUR-TEAM 1.00000 0.03396
3 CarbonSiliconAI 0.97959 0.06710
Locomotion
Locomotion
Phase 1:
Rank Participant team Time (↓) Score (↑) Points (↑)
1 GaitNet 3.20980 0.83951 0.97959
2 MSKBioDyn 10.19122 0.49044 0.61224
3 Lattice 12.61408 0.36930 0.48980
Phase 2:
Rank Participant team Score (↑) Effort (↓)
1 Lattice 0.34343 0.05220
2 GaitNet 0.30303 0.05506
3 CarbonSiliconAI 0.20202 0.07620
Phase 2:
Rank Participant team Time (↓) Score (↑) Points (↑)
1 GaitNet 11.06697 0.66613 0.63636
2 MSKBioDyn 11.08687 0.23651 0.20202
3 CarbonSiliconAI 13.56232 0.17307 0.09091
Sponsors
Sponsors
Citations
Citations
@Misc{MyoChallenge2023, author = {Vittorio Caggiano, Huawei Wang, Guillaume Durandau, Seungmoon Song, Chun Kwang Tan , Cameron Berg, Pierre Schumacher, Massimo Sartori, Vikash Kumar}, title = {MyoChallenge 23: Towards Human-Level Dexterity and Agility}, howpublished = {\url{ https://sites.google.com/view/myosuite/myochallenge/myochallenge-2023 }}, year = {2023}}
Page updated
Google Sites
Report abuse