Zunzhi You

Hi! I am Zunzhi You (尤尊智). You can call me Dazhi (/dɑ:dʒ-/).

I am a first-year PhD student at USYD, supervised by Dr Chang Xu. My research interests center around the robustness and interpretability of machine learning models in computer vision applications.

I recently graduated from the School of Computer Science and Engineering, Sun Yat-sen University as an outstanding graduate, advised by Dr Guanbin Li. I was also fortunate to work with Dr Wei-Chen Chiu and Dr Yi-Hsuan Tsai since my visit to National Chiao Tung University (now NYCU). More recently, I had a wonderful academic visit at the SNU CVLab, advised by Professor Bohyung Han.

Email  /  CV  /  GitHub  /  Zhihu

We show that a simple variant of mask image modeling can be used to provide adversarial robustness in addition to pretrained features.

We quantify the interpretability of deep networks for Monocular Depth Estimation and further propose a method to make them interpretable.

We propose a method that colorizes a depth map with (photo-)realistic appearance.

First, we see the derivation of how to predict a target using Gaussian Process Regression in supervised learning. Then, we implement a toy example. Lastly, we understand the algorithm by figuring out the key assumption.