Collecting Movement Data and Predicting Surgical Outcomes in the Age of Deep Learning
Recent achievements in machine learning are subverting foundations of many research disciplines. In this presentation, I will show how these developments affect collection, modeling and prediction of human gait kinematics, enabling unprecedented scale of research and applications. I will focus on our novel technique for predicting progression trajectories of pathologic gait kinematics from sparse observations, using matrix completion techniques. I will present how we collect movement data with equipment 100x cheaper than usual and how we model movement using reinforcement learning, leveraging large computational resources and domain knowledge embedded in simulation software.
Łukasz Kidziński is a researcher in the Mobilize Center at Stanford University, working on the intersection of computer science, statistics and biomechanics. Previously a data scientist in the CHILI group, Computer-Human Interaction in Learning and Instruction, at the EPFL. His main interests include computational methods in biomedical data, including applications of machine learning, data mining, big data, time series analysis and statistics.