Thomas Ploetz

Georgia Institute of Technology

Title: Computational Behavior Analysis for Wellbeing Assessments — Smart Entities (Devices, Algorithms, Humans) are Pushing the Boundaries of Digital Health.

Abstract: Many aspects of both mental and physical wellbeing are linked to a human’s behavior and activities, i.e., what a person is doing and when, and how this changes over time. As such, human activity recognition using wearable and other pervasive sensing methods (HAR) is a pillar of mobile and ubiquitous computing based, out-of-clinic wellbeing assessments in various shapes and forms. Alas, recognizing what a person does is—to this day—a non-trivial endeavor due to the underlying analysis of multi-variate, noisy sensor data, which constitutes a formidable machine learning problem that the research community has been tackling. In this talk I will first summarize what makes HAR a hard machine learning problem and then discuss how recent technological breakthroughs can help to get closer to solving it. I will focus on opportunities to overcome data scarcity through cross-modality transfer, and self-supervised learning. I will then delve into some of the wellbeing assessments that my lab has been doing in areas such as Parkinson’s assessment, automated analysis of eating behavior as proxy for mental wellbeing assessments, and recent work in behavior assessments for people with mild cognitive impairments. Finally, I will explore some more foundational thoughts on how we (should) characterize activities and behaviors and how language learning may help us to develop a more systematic and thus robust and generalizable understanding of human activities. I argue that human activities follow concepts that are similar to those of natural languages and as such these can be learnt, which will lead to improved modeling as well as analysis — as it is of importance, for example, for many medical applications.

Bio: Thomas Ploetz is a Computer Scientist with expertise and almost two decades of experience in Pattern Recognition and Machine Learning research (PhD from Bielefeld University, Germany). He works as a Professor of Computing at the School of Interactive Computing at the Georgia Institute of Technology in Atlanta, USA. His research agenda focuses on applied machine learning, that is developing systems and innovative sensor data analysis methods for real world applications. Primary application domain for his work is computational behavior analysis where he develops methods for automated and objective behavior assessments in naturalistic environments, thereby making opportunistic use of ubiquitous and wearable sensing methods. Main driving functions for his work are "in the wild" deployments and as such the development of systems and methods that have a real impact on people's lives. Thomas has been very active in the mobile and ubiquitous, including wearable computing community. He is co-editor in chief of the Proc. of the ACM on Interactive, Mobile, Wearable, and Ubiquitous computing technology (IMWUT), has twice been co-chair of the technical program committee of the International Symposium on Wearable Computing (ISWC), and was general co-chair of the 2022 Int. Joint Conf. On Pervasive and Ubiquitous Computing (Ubicomp). Thomas is a Distinguished Member of the ACM. More info here.