Sachini Piyoni Ekanayake is a Ph.D. student in the Electrical and Computer Engineering Department at the University at Albany, State University of New York, NY, USA, working with Dr. Daphney-Stavroula Zois. She received her B.Sc. degree in Electrical and Electronic Engineering from the University of Peradeniya, Sri Lanka, in 2017. She was a machine learning fellow intern at GE Research, Niskayuna, NY, USA, in the summer of 2022. Her research interests include machine learning and statistical signal processing applications in Cyber-Physical Systems.
Knowledge of human context information (e.g., affective state and physical activity) offers valuable insights into human behavior. It also enables the design of assistive technologies through accurate predictions for decision-making in real-time. However, human context variables are not directly observable but are observed through noisy and costly features and may vary across individuals. To address these concerns, I have created an experimental platform that communicates and collects sensor data, and designed algorithms to estimate human context variables and their relationships from such data. Specifically, I have developed an algorithm for learning relationships between Bayesian network variables. I have also devised algorithms that jointly select the most informative features and the Bayesian network variable values. To further enhance classification accuracy, I have also devised a dynamic feature and classifier selection strategy. The instance-wise nature of the proposed algorithms provides valuable insights into the decision-making process, enabling interpretability.