Tianzhe Chu

I’m a rising senior-year undergraduate student major in Computer Science at ShanghaiTech University, with a wonderful year(22-23) spent in UC Berkeley. I'm from Suzhou(Soo Chow), China.

Email  /  Resume  /  Classes  /  Github  /  Twitter

profile photo
Research

I am currently an undergraduate researcher in Berkeley Artificial Intelligence Research Lab(BAIR) advised by Prof. Yi Ma, working closely with Tianjiao Ding from JHU and Peter Tong from NYU Courant.

My research interest broadly lies in representation learning. Currently, I'm working on unsupervised/self-supervised learning, interpretable learning architectures and vision-language models. Besides, I'm open to any interesting research topics related to intelligence.

Feel free to reach out to talk about research, music, cooking, pokémon, etc.

News

May 2023: Gonna leave lovely Berkeley (as well as U.S.), finishing 7 tech courses and a few interesting research projects.

Publications & Preprints (* means equal contribution)
PontTuset Emergence of Segmentation with Minimalistic White-Box Transformers
Yaodong Yu*, Tianzhe Chu*, Shengbang Tong, Ziyang Wu, Druv Pai, Sam Buchanan, Yi Ma
Under Review
demo / project page / code / arxiv

The white-box transformer leads to the emergence of segmentation properties in the network's self-attention maps, solely through a minimalistic supervised training recipe.

PontTuset Image Clustering via the Principle of Rate Reduction in the Age of Pretrained Models
Tianzhe Chu*, Shengbang Tong*, Tianjiao Ding*, Xili Dai, Benjamin D. Haeffele, René Vidal, Yi Ma
Under Review
project page / code / arxiv

This paper proposes a novel image clustering pipeline that integrates pre-trained models and rate reduction, enhancing clustering accuracy and introducing an effective self-labeling algorithm for unlabeled datasets at scale.

PontTuset White-Box Transformers via Sparse Rate Reduction
Yaodong Yu, Sam Buchanan, Druv Pai, Tianzhe Chu, Ziyang Wu, Shengbang Tong, Benjamin D. Haeffele, Yi Ma
Accepted by NIPS 2023
code / arxiv

We develop white-box transformer-like deep network architectures which are mathematically interpretable and achieve performance very close to ViT.