Research Assistant at Scalable Analytics Institute, UCLA 2019-2023

Several projects we work on:

Twitter ideology prediction with graph neural network based models, an encoder-decoder structure is used to fully utilize the graph information through semi-supervised learning.

Multimodal fake news detection, combining both the graph information and text information in a social network to achieve the best performance.

Active learning on graphs, boosting the performance of current GNN models using active learning.

Software Engineer Intern at Aurora 2022

Built novel deep learning scene classifiers for autonomous vehicle perception using pytorch framework. Achieved 99% precision on a group of scene classification tasks.

Machine Learning Research Intern at Alibaba 2021

Developed a reinforcement learning model which successfully improved the clean energy consumption in a power grid by more than 50%.

Our model won the First Award in the Smart Grid Unit Commitment Competition organized by State Grid.

Developed time series forecasting algorithms for short term fluctuation of solar energy and wind energy output.

Remote Research Intern 2020-2021

Developed novel multi-agent reinforcement learning algorithms for multi-robot coordination problems, under the supervision of Professor Jun Wang from University College London. Those algorithms could help achieve basic coordination between a group of autonomous vehicles.

Risk Advisory Intern at Deloitte 2019

Developed a machine learning based portfolio risk assessment model. Successfully detected the portfolios that exceed the risk limit with 90% accuracy and 97% recall.