Testimonials

Testimonials
2024 Summer KAIST CAMPUS Asia Research-oriented Program_Shi Zhiwen from Tsinghua University
2024-12-11

Shi Zhiwen from Tsinghua University


In April 2024, I found a notice about the Campus Asia exchange program on the school information website. I hoped to enrich my learning and cultural experience, so I submitted the application materials to the program committee. Fortunately, I have obtained the qualification to participate in the exchange program and look forward to starting my trip to Korea in August. 


There are ten exchange students in the KAIST Asia Campus Program, three are Japanese students from Tokyo Tech, and the remaining seven are Chinese students from Tsinghua University. We attended classes, had meals, and traveled together, establishing a deep friendship within a month. It's inevitable to be reluctant to say goodbye when leaving KAIST, so we make an appointment to meet again in Tokyo and Beijing next time. As a student of Shenzhen Graduate School of THU, I hope to have the opportunity to reunite with friends in Beijing later, and I also welcome friends to come to Shenzhen and meet at the beautiful THU campus in the south of China.


Thanks to the kind coordinators Soohee Lim, Yuna Kim, and Prof. Kim in the KAIST Global office, they prepared the warm orientation & ending ceremony and planned the wonderful school trip to Jeonju! We also took group photos together with Hanbok in front of Gyeonggijeon Palace. Also, the buddy students Edward Kim and Minjeong Lee helped us a lot when we lived in the dormitory of KAIST and gave us a detailed campus tour in the first week!


My program is a research-oriented program, so I chose a laboratory in KAIST to do some research in one month. The Neuro-Machine Augmented Intelligence Laboratory(NMAIL), belongs to the School of Computing at KAIST, and my supervisor is Prof. Sungho Jo. At the first meeting, the professor was very enthusiastic and shared some advice on career choices with me, taking me to visit three labs with different directions, mainly including neural enhancement, machine learning, and soft robotics, which is compatible with my professional background of BME in THU. My research topic is understanding Generative Models and my PhD student mentor in NMAIL is Sejoon Huh. He shared relevant literature and code with me, and arranged offline tutorials 2-3 times a week to solve problems, providing me with feasible solutions. Throughout my research study within one month, I learned basic GAN and VAE models and tried to fine-tune the diffusion models. Due to the time limits, it has not been better combined with medical images, but it has also inspired my research ideas, which is helpful for subsequent research topics. After this summer research, I also had the idea of continuing my PhD study abroad.


I also experienced the local festivals and traditional culture in Korea. August 14th is the Korean National Holiday. It was Thursday, and everyone had one day off and walked on the streets to fully enjoy the festival celebrations and atmosphere. At the same time, there is also a Zero Clock Celebration event being held in Daejeon, which is divided into three stages: Daejeon's past, present, and future, on the main road near the Daejeon railway station. We watched the band competition at the celebration event in the evening. There is a beautiful sky road here, with LED screens covering the ceiling built along the street. I also love the campus environment at KAIST very much. On many clear summer afternoons, I can always see the beautiful sky, clouds, and sunset.


In the one-month study in KAIST, I can gradually build the knowledge system for generative models and overview the raw materials of model research. I can catch up with the new direction for model developments and updates, and get to be more familiar with the three classical architectures, including GANs, VAEs, and Diffusion Models. I have the opportunity to enhance my coding ability for program practice, including the usage of Pytorch, Jupyter Notebook, Github, and Hugging Face. Meanwhile, it involves math theory study and formulation derivation during the process of model adaptation and fine-tuning. I experience the culture and research environment, such as the research discussions (thanks to my mentor Sejoon Huh :), PhD student in NMAIL) and some cultural communications (friends in China, Korea, and Japan, Korean Language Courses). I really enjoy the exchange experience at KAIST and have a lot of unforgettable memories. Hope my friends are all good anywhere around the world and we can meet again someday in the future!