Real-time bidding

- Developed a technique to enhance the sharing efficiency of complementary-label learning algorithms, leading to significant improvement in the accuracy of those algorithms.

[ Graduate Research Assistant at Computational Learning Lab ]Advised by Prof. Hsuan-Tien LinAddressed the problem of complementary label learning. Proposed a novel perspective that (a) led to a learning framework with theoretical guarantees, (b) enabled any ordinary classifier to learn from complementary labels, and (c) explained previous methods as special cases.[Outcome: A paper accepted by PAKDD 2023 and awarded best paper runner-up.]

[ Head TA in Machine Learning, Fall 2021 ][ TA in Data Structures and Algorithms, Spring 2021 ][ TA in Machine Learning Foundations, Fall 2020, Fall 2018 ][ TA in Machine Learning Techniques, Fall 2020 ][ TA in Machine Learning, Spring 2019 ] - Evaluated homework and held TA hours to ensure students understood the course material.

- Built and enhanced the recommendation systems for LINE shopping. - Improved jieba to be a more accurate and efficient Chinese tokenizer for analyzing NLP data in LINE shopping.- Shared technical insights on LINE Taiwan's engineering blog.

- Sought trading signals and designed trading strategies (alpha) in the US stock market with mathematical and statistical models.- Received Best Alpha Quality Award and Best Alpha Combo Award for building the alpha pool with highest average quality and highest performance of combination among all interns.