2023-04-18 •

COURSE PROJECT

CSC2516 Neural Networks and Deep Learning | UoT | March 2023 - April 2023

Project Title: CLIP-guided Zero-Shot Text-to-Image Generation

Proposed a method to perfrom zero-shot text-to-image generation based on CLIP guidance.

Implemented the proposed method with PyTorch.

Conducted experiments on CUB-200 dataset and our method outperformed all the baseline methods.

Wrote a report in a conference format with teamates.[report][project]

CSC2515 Introduction to Machine Learning | UoT | Nov 2022 - Dec 2022

Project Title: Learning Disentangled Features for Domain Generalization

Proposed novel methods to disentangle the feature by swapping.

Implemented the initial version of proposed method with PyTorch.

Conducted experiments on Digit-DG dataset and our method outperformed all the baseline methods.

Wrote a report in a conference format with teamates.[report][project]

CSC2529 Computational Imaging | UoT | Nov 2022 - Dec 2022

Project Title: Can Diffusion Model Generalize Well in Image Super Resolution with Limited Fine-Tuning?

Fine-tuned a pretrained SR3 model with limited update steps, amount of data and time steps on a new data domain which are unseen during pretraining.

Verified the pretrained model’s generalization ability qualitatively and quantitatively in terms of PSNR, SSIM and LPIPS.

Wrote a report in a conference format and made a poster presentation.[report]

Introduction to Artificial Intelligence | BIT | Dec 2020 - Jan 2021

Implemented CNNs for checkerboard recognition with Pytorch.

Implemented gobang AI algorithms based on the minimax search strategies, evolutionary algorithms, and deep Q-Learning.

Wrote a report to analyze the results of the experiments above. The gobang AI based on minimax search and Deep Q-Learning approaches achieved good performance which is comparable to humans’.

Course group of Assembly Language and Interface Technology | BIT | Group leader | May 2021 - May 2021

Implemented a rhythm game only with x86 assembly language.

Assigned tasks to other group members and monitored their progress, helped them to solve difficult problems and coordinated with them.

Our group received A for this course project.

This project can be found here: https://github.com/107dot25/Ark-Beat

Course group of Computer Science fundamental practice | BIT | Group leader | Sep 2020 - Sep 2020

Implemented a software to simulate a simple telemedicine monitoring system with QT.

Assigned tasks to other group members and monitored their progress, helped them to solve difficult problems and coordinated with them.

Our group is the first to complete the task, and I received 99 points for this course.