Tien-Tuan DAO seminar on biomechanical issues

Title of presentation

Neuromusculoskeletal modeling coupled with reinforcement learning for learning human locomotion: From walking to age-related falls and recovery.

Abstract

Modern artificial intelligence (AI) has been the subject of intensive research in many fields ranging from computer vision to precision medicine. Different learning strategies such as deep learning, transfer learning or reinforcement learning have been developed to move from applied artificial intelligence to the level of general artificial intelligence. In this new era, the field of biomechanics has also been metamorphosed with new AI-driven models and decision aids.

In this talk, we will discuss our recent achievements in using reinforcement learning coupled with neuromusculoskeletal modeling to explore human locomotion. First, novel bioinspired reward remodeling strategies will be presented to elucidate muscle coordination during different gait models. Secondly, age-related falls and recovery will be addressed to explore human locomotion strategies and motor learning for an aging population using transfer tests and coupled simulations. In fact, the coupling between neuromusculoskeletal modeling and reinforcement learning opens up new avenues for patient-specific functional rehabilitation, fall prediction and associated preventive strategies.

Brief biography

Tien-Tuan Dao is a full professor in the Mechanical Engineering Department at Centrale Lille Institut (France). He coordinates the Soft Tissue Biomechanics group of the LaMcube laboratory (UMR CNRS 9013 Laboratoire de Mécanique, Multiphysique, Multiéchelle). Dao focuses on the study of pathogenesis and salutogenesis processes of complex bio-systems (face, lumbar spine, pelvis, lower limbs) towards new predictive and preventive care.