Machine Learning Intern at Apollo, Baidu USA (Sept. 2021 - Dec. 2021)
Interaction in Autonomous Driving
- Propose and implement the rule-based control point decider and gap decider, when the autonomous vehicle is in the unprotected left turn scenario at the intersection. They increases the efficiency of unprotected left turn.
- Implement a prototype reinforcement learning model (DQN) by Libtorch to determine the control point and gap to pass. Prove the feasibility of RL in unprotected left turn.
Research Assistant at Northeastern University (Sept. 2018 - Now)
Resource Allocation
- Experimental design networks for serving heterogeneous learners
- Distributed, adaptive, algorithms optimizing cache decisions
- Joint optimization of caching and routing
- Cache networks incorporating queuing models
- Fair cache networks
Machine Learning
- Experimental design networks for serving heterogeneous learners
- Multimodal learning for vehicle detection