Understanding Deep Generative Models
The aim of our project is to develop systems in which users can define constraints for these dimensions in a declarative manner, so that a system can automatically suggest a suitable sequence of ML processes. These systems can address the constraints at both the data and model levels. For the data level, we have developed systems that work with feature selection and construction. For the model level, we have developed systems that solve the problem as hyperparameter optimization.
Insight into Deep Generative Models with Explainable AI
The just completed agility project "Understanding Deep Generative Models" focused on making ML models more transparent, in order to uncover hidden flaws in them, a research area known as Explainable AI.
Research Group Lead
Fellow
Director
Postdoctoral Researcher