Soupy twist!

I am Ching Lam, a final year undergraduate from CUHK studying AI Engineering, with a double minor in English and German. I am applying for PhD positions in the 2024 fall entry season, to better understand robustness in computer vision (CV) and machine learning (ML).

I’ve been incredibly fortunate to spend time with amazing people at exciting places. From Feb - Jun 2024, I will be interning with Serge Belongie at SØ(3) / Belongie Lab at the University of Copenhagen. 2023 was a particularly formative year, during which I completed a May - Aug internship at MILA with Yann Dauphin and Aaron Courville; also, an extended internship with Wieland Brendel at Tübingen’s Max Planck Institute for Intelligent Systems (MPI-IS) (Jan - Apr, 2023 and Sep 2023 - Jan 2024). In 2022, I did a remote, 6-month internship with Jiajun Wu at the Stanford AI Lab (SAIL); I also started as a research assistant with CUHK’s Farzan Farnia in the same year. From 2020 - 2021, I was a research intern at NVIDIA; from 2019 - 2022, I worked with my first-ever mentor, Hongsheng Li, at CUHK’s Multimedia Laboratory (MMLab).

My idol in life is Hugh Laurie and nothing brings me more joy than A Bit of Fry and Laurie.

Academic Interests

I am fascinated by the intersections between ML and CV, including reconciling the various definitions of “robustness”. I want to tackle fundamental problems such as whether deep learning methods are generalizable learners, how we should define and design for explainable/interpretable models, and how to better understand learning dynamics under noisy, multi-agent settings (e.g. label noise, distributional shifts, adversarial training, knowledge distillation, federated settings). I deeply enjoy both theoretical/empirical research; I strive to connect theory and practice, by building theoretically sound and practically grounded models.