š± About Me
Hi! I am an senior student majored in computer science at Shanghai Jiao Tong University (SJTU), China. Currently, I am engaged in research work in MultiMedia Lab under the supervision of Professor Xiaohong Liu. My research interests mainly focuses on computer vision and artificial intelligence generated content.
š News
Iām currently (from July 2024) an exchange visitor for research purposes in Vision and Learning Lab at University of California Merced, under the supervision of Professor Ming-Hsuan Yang.
š Educations
- 2021.09 - now, Shanghai Jiao Tong University, Shanghai, China. (Bachelor of Science)
- 2018.09 - 2021.06, Zhejiang Ruian High School, Zhejiang, China.
š Honors and Awards
- 2021, 2022, 2023 Zhiyuan Honors Scholarship
- 2022, 2023 Undergraduate Class B Scholarship
- 2022.2 MCM/ICM Class M Award
š Course Projects
Using information theoretic metrics to study the importance of individual neurons in DNNs
Information Theory course project: In this paper, I use information theoretic metrics for node pruning to learn the importance of individual neurons at different levels in the whole DNN. Entropy, Mutual information and KL-Selectivity are used to determine the order of ablation.
Advisor: Prof. Fan Cheng
ICE2601: Information Theory
SJTU, 2023 Spring
For Fashion-MNIST clothing classfication, I designed a deep learning network on my own named FashionNet, which shows excellent performance on understanding and representing the input image and classfication task. And also, I employed a VAE to deal with image reconstruction task.
Advisor: Prof. Quanshi Zhang
CS3308: Machine Learning
SJTU, 2023 Spring
This is the course project I designed on my own for CS2107 where I served as TA, which is basically generated from Graphics And Mixed Environemnt Seminar, Lingqi Yan, UCSB, but chooses to use a more modern programming language, i.e. Rust. The project includes 3 LABs, simply allows students to learn the basics of rasterization in graphics and, most importantly, to have fun.
Advisor: Prof. Qingsheng Ren
CS2107: Programming languages and Data structures III
SJTU, 2023 Summer
We introduce a novel bootstrapping approach for training generative models. Specifically, we construct synthetic datasets by combining generated samples from previous iterations with real data. Our empirical results demonstrate the potential for performance improvement through bootstrapping diffusion models. By recycling samples over successive generations, this technique reduces the dependence on large curated datasets while producing varied outputs.
Advisor: Prof. Jianfu Zhang
CS3967: Image Processing and Computer Vision
SJTU, 2023 Fall
š» Experiences
- 2023.06 - 2023.07, Serve as a teaching assistant for CS2107.
- 2023.06 - 2023.09, BASIC Lab, under the supervision of Prof. Qiang Yin.
- 2023.10 - now, MultiMedia Lab, under the supervision of Prof. Xiaohong Liu.