Some specific research topics that he is interested in and currently working on: 

• Topic modeling: He is interested in developing various topic models (either probablistic or deep learning based) that can explore different discourse structures and other meta information associated with natural language text.
• Active Learning: Active Learning is an approach to speeding up learning by judiciously selecting data to be annotated. His group has been working on developing effective and efficient acquisition functions that stay in the core of active learning.
• Uncertainty Estimation: Uncertainty estimation is essential to make ML/DL trustworthy in real-world applications, particulalry the stafy-critical applications (e.g., AI-aided diagnosis). He is inerested in equiping ML/DL models with capability of estimating uncertainty in their predictions.
• Applications in Public Health: He has been working closely with Medical experts on developing and deploying cutting-edge ML/DL technologies for AI enabled public health.