Talk on Applied Machine Learning, by Yongqi Dong

Yongqi Dong, a researcher at University of Waterloo visited Certus on June 17/18, and gave a talk on applied machine learning in shared mobility.

Abstract: In this presentation, I will share my exploration in ML and data-driven research and my general ideas about Validation intelligence for Autonomous Systems, especially ML-based systems.

Firstly, I will talk about Data-Driven Research in Shared Mobility, focusing on how to modeling shared mobility problem from a spatial-temporal aspect. (An Empirical Study on Travel Patterns of Internet Based Ride-Sharing published on Transportation Research Part C).

Then, I will introduce my exploration in DL for prediction and classification problems, using my participated Kaggle Competitions. Also, I will cover a little about Deep Reinforcement learning for controlling problems. A proposed shared mobility solution which combines DL for prediction and DRL for controlling will also be introduced.

Finally, I will share my general ideas about testing ML-based systems, following the three main challenges, corresponding solutions, as well as potential research directions.

 

Bio: Yongqi Dong, Master of Control Science and Engineering, now serves a graduate research assistant at the University of Waterloo. Before coming to Canada, Yongqi Dong got his Master degree at Tsinghua University and Bachelor degree in Telecommunication at Beijing Jiaotong University. Yongqi Dong possesses a Master Minor in Big data and Machine Learning. His research mainly falls into data-driven interdisciplinary research using applied ML methods, especially in shared mobility domain.