Focus on developing new recommender system with cold start item.
Explore shared network structure and different attention mechanisms to improve the performance of item cold start recommender system.
Research Assistant at Laboratory of Machine Learning & Health Informatics, UCONN 01/2016-present
Develop new machine learning algorithms, such as for sparse learning problem, optimizers for deep learning applications, differential private (DP) algorithms and federated learning algorithms.
Collaborate across labs to identify new composite traits for alcohol dependence.
Graduate Assistant at Eversource Energy Center, UCONN 05/2017-present
Implemented weather forecasting ( WRF 3.7, WRF 3.8 and ICLAMS systems ) , Flooding forecasting ( CREST system ) and machine learning models for power outage predictions for Eversource Energy ( CT, MA, NH areas ) and The United Illuminating Company (UI) on the HPC resources at UConn.
Proposed new machine learning and deep learning methods to improve the performance of existing operational system, such as LSTM model to capture the sequence information in weather.
