Celina Chen

Trustworthy AI · University of Toronto

Intelligence,
worthy of trust.

I'm Celina, an M.A.Sc./PhD student at the University of Toronto. I study how to make machine learning models private, secure and robust, and how those guarantees trade off with what the models can do.

See my work CV
Portrait of Celina Chen

01Education

Sep 2026 – Present
M.A.Sc./PhD, Electrical & Computer EngineeringUniversity of Toronto · Advised by Prof. Nicolas Papernot
Sep 2023 – May 2026
B.Sc. Honours, Computer Science SpecialistUniversity of Toronto · Full ride as a Schulich Leader Scholar

02Two questions I keep returning to

Agent security

As AI agents take on more of our work, how do we make sure every action they take keeps our systems safe?

Interpretability

How do models understand what they learn? Their internal representations may hold the key to core problems like adversarial robustness and continual learning.

Looking ahead

I believe continual learning is the most important open problem in AI: understanding how models keep learning after deployment, and catching it when that learning goes wrong. My work is a step toward it.

03Selected publications

  1. arXiv 2026 Differentially private mixing of public datasets improves private learning

    Yufei Chen, Tejumade Afonja, Anvith Thudi, Nicolas Papernot

    arXiv Google Scholar Poster

Always happy to talk about trustworthy AI, research ideas, or collaborations.

Write to me