Research, design and implement models for Ad bid prediction, optimization and analysis• Project on action prediction models with historical impression data• Project on joint budget optimization for maximal action rates• Project on rare events prediction• Project on user sampling for campaign performance predictions

* NSF funded research Intelligent Clinical Decision Support with Probabilistic and Temporal EHR Modeling • Project on exponential family models for mining count data from EHR (IEEE BIBM 2017) • Project on pattern discovery and event prediction in structured sequence data (AAAI 2016) > Proposed the syntax and semantics of the first continuous-time probabilistic logic model which allows anytime sequential event prediction for structured sequence data > Proposed an efficient relational gradient boosting approach for learning the proposed model and experimented on propositional as well as relational sequence data • Project on developing dynamic probabilistic models for pattern discovery and sequential events prediction for longitudinal clinical data in a Cardiovascular disease study (AIME 2015) • Project on probabilistic causal network learning for causal dependency discoveries in medical data • Project on hybrid probabilistic logic models for heterogeneous relational data* Research on advice-based statistical relational learning • Project on cost-sensitive learning for class-imbalanced structured Data (ICDM 2014) > Proposed a cost-sensitive relational gradient boosting approach to allow the trade-off between false positive rate and false negative rate during the learning process • Project on advice-based statistical relational learning for sequential event prediction

Project on personalized video recommendation systems

Project on semi-supervised model adaptation, online learning with concept drift

Project on building large-scale hybrid job recommender systems with probabilistic logic models

* Research on knowledge-based learning for high-dimensional sparse data (ECMLPKDD 2013) > Proposed a way to incorporate domain knowledge on qualitative constraints and independence of causal influence into Bayesian network learning * Project on analysis of heart rate variability in time and frequency domains

Analyze and improve the circuit design of commercial inverter air conditioner

• Developed algorithms for auto-detection of QT intervals in ECG signals• Implemented the algorithm in the embedded system of ARM7 for a portable ECG monitor