Major:

• Mathematics (Applied Mathematics Track)
• Additional Major in Computer Science

Awards and Qualifications:

• The S.S. Chern Class of elite and talented students in Mathematics - Department of Mathematics, HKUST
• University Scholarship Scheme for Continuing Undergraduate Students 2019-2020
• Dean’s List of the School of Science for the Fall Term of 2018-2019 - School of Science

Relavent Courses:

• 2020Fall: MATH3312 - Numerical Analysis (A+), HUMA1000G - Cultures and Values: Love, Death and Human Nature (A), COMP5331 - Knowledge Discovery in Databases.
• 2020Spring: MATH2431 – Honors Probability (A+), COMP3111 - Software Engineering
• 2019Fall: MATH2131 - Honors in Linear and Abstract Algebra I (A), MATH3033 - Real Analysis, COMP3711H - Honors Design and Analysis of Algorithms, COMP2012 - Object-Oriented Programming and Data Structures, COMP1943 - Creative Sound Design (A).
• 2019Summer: MATH2411 - Applied Statistics (A+).
• 2019Spring: MATH1024 - Honors Calculus II, MATH2043 - Honors Mathematical Analysis, COMP2011 - Programming with C++.
• 2018Fall: COMP2711H - Honors Discrete Mathematical Tools for Computer Science (A+), MATH1023 - Honors Calculus I (A).

Specialization Certificate

Course Certificates Completed:

• Structuring Machine Learning Projects

• Sequence Models

• Neural Networks and Deep Learning

• Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization

• Convolutional Neural Networks

• Supervisor: Xiaojuan Ma
• Description: Nowadays, AI has been applied in many interactive systems, increasing people’s subjective well-being in various domains. For example, recommendations on happiness-leading activities can make people happy. The goal of our proposed system, HappyFinder, is to increase variety and enable easy personalization of the recommendations of happiness leading activities, which can most effectively enhance happiness. We analyze happiness moments from HappyDB, extract happy activities and conduct data augmentation. In the recommender system, we measure similarities between happy activities with a commonsense mediated method and recommend activities according to users' preferences through their interaction with the system.

• Supervisor: Jian-Feng Cai
• Description: This is an independent study on matrix decompositions using randomized algorithm and its applications. I read relevant papers on randomized methods for matrix approximation, learn applications of the methods on the latest technologies, and try to explore new methods to solve real-world application problems with the guide of the instructor.

• Supervisor: Han Zhang
• Description: This research project aims to use graph convolutional networks (GCN) to learn representations of survey respondents as well as their answers. In this project, we try to come up with different methods to measure the similarities between users.

• Supervisor: Desmond Yau-chat Tsoi
• Description: In this project, we explored deep learning-based document restoration approaches to correct geometric and shading artifacts of camera-imaged paintings. In the spring semester, we focus on the construction of 3D meshes and the generation of the training datasets. In the summer semester, we explore some others' methods of correcting images of distorted documents and apply those methods to the dataset with distorted painting images.

• Supervisor: Jian-Feng Cai
• Description: We learn the Variational Autoencoder from two different perspective: the neural net perspective and the probability model perspective. In this research opportunity, we tried to explain the probability model in a more mathematical way and compared the two perspectives.