RT Journal Article
SR Electronic
T1 Acquiring musculoskeletal skills with curriculum-based reinforcement learning
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 2024.01.24.577123
DO 10.1101/2024.01.24.577123
A1 Chiappa, Alberto Silvio
A1 Tano, Pablo
A1 Patel, Nisheet
A1 Ingster, Abigaïl
A1 Pouget, Alexandre
A1 Mathis, Alexander
YR 2024
UL http://biorxiv.org/content/early/2024/01/25/2024.01.24.577123.abstract
AB Efficient, physiologically-detailed musculoskeletal simulators and powerful learning algorithms provide new computational tools to tackle the grand challenge of understanding biological motor control. Our winning solution for the first NeurIPS MyoChallenge leverages an approach mirroring human learning and showcases reinforcement and curriculum learning as mechanisms to find motor control policies in complex object manipulation tasks. Analyzing the policy against data from human subjects reveals insights into efficient control of complex biological systems. Overall, our work highlights the new possibilities emerging at the interface of musculoskeletal physics engines, reinforcement learning and neuroscience.Competing Interest StatementThe authors have declared no competing interest.