Ruozhu Yao Email & Phone Number

Leiden, South Holland, Netherlands

Netherlands

During the 9-month Master's Graduation project, I conducted research on the topic of "How to make accurate predictions in stream data with the existence of concept drifts".I read over 20 papers about concept drifts and the existing solutions, and I tried 4 solutions found on GitHub; this led to the design of an online retraining framework that combines a simple neural network for online next activity prediction and the "Online continual Learning Algorithm". The target framework is able to make accurate predictions efficiently for stream data with drifts. The implementation and details of the target framework can be found in my thesis (not published yet).For my graduation project, I got a grade of 8 for my excellent performance during the defense and the good-quality report.

Amsterdam, North Holland, Netherlands

During the 1.5-month summer internship, I worked on the cold start problem in the advertisement recommender system.With a team of 5, we analyzed the problem and found possible solutions. We had daily meetings during which we presented findings to 2 experienced data scientist coaches. We also held daily working sessions when another teammate and I taught other teammates without technical backgrounds Python and Github; this led to a presentation providing an extensive analysis of the problem and one possible solution, and codes with basic implementation of the proposed solution. On the final presentation day, I was awarded the "Best Presenter" out of over forty people for my vivid presentation of the problem and the corresponding solution.

Shenzhen, Guangdong, China

During a one-year internship, I worked on the research project ‘How to Evaluate and Boost the Robustness of a Neural Network.’ I conducted extensive research, leading to the development of two algorithms ‘Statistical Classification Robustness Revisit (SCR2)’ and ‘Adaptive Retraining (ART).’ SCR2 can calculate the network's robustness for evaluation, and ART was instrumental in enhancing model stability, as evidenced in benchmark tests. These algorithms, detailed in a published paper that was cited 2 times, were integrated into the company’s deep learning platform. My work was recognized as one of the best intern projects for its innovative approach and impact on neural network optimization.

Guangzhou, Guangdong, China

Master Of Science - Ms, Data Science And Artificial Intelligence, 8

Activities and Societies: Play Squash and dance, active members in Footloose and Confluente,

Bachelor Of Science - Bs, Applied Mathematics