Isay Katsman

I will be joining Yale University in Fall 2022 to do a PhD in applied mathematics under the supervision of Prof. Anna Gilbert! Currenty, I am a Computer Science MS student at Cornell University advised by Prof. Chris de Sa. I finished a Math and CS double major as an undergrad at Cornell in 2020, advised by Prof. Serge Belongie.

My research focuses on geometric deep learning and in particular on learning generative models over manifolds (with especially close consideration given to symmetry and equivariance). I also work on exploring flaws in deep learning models and ways to correct them through my work on adversarial examples.

Email: [Google Scholar] [LinkedIn] [Github]


  • April 2022: I will be doing a PhD at Yale starting Fall 2022!
  • August 2021: One paper accepted to NeurIPS 2021.
  • August 2020: One paper accepted to NeurIPS 2020.
  • June 2020: One paper accepted to ICML 2020.


Differential Geometry in Machine Learning

(NeurIPS 2021) Equivariant Manifold Flows [Arxiv]

Isay Katsman*, Aaron Lou*, Derek Lim*, Qingxuan Jiang, Ser-Nam Lim, Christopher De Sa

(NeurIPS 2020) Neural Manifold Ordinary Differential Equations [Arxiv]

Aaron Lou*, Derek Lim*, Isay Katsman*, Leo Huang*, Qingxuan Jiang, Ser-Nam Lim, Christopher De Sa

(ICML 2020) Differentiating through the Fr├ęchet Mean [Arxiv]

Aaron Lou*, Isay Katsman*, Qingxuan Jiang*, Ser-Nam Lim, Serge Belongie, Christopher De Sa

Adversarial Examples

(ICCV 2019) Enhancing Adversarial Example Transferability with an Intermediate Level Attack [Arxiv]

Qian Huang*, Isay Katsman*, Horace He*, Zeqi Gu*, Serge Belongie, Ser-Nam Lim

(CVPR 2018) Generative Adversarial Perturbations [Link] [Code]

Omid Poursaeed, Isay Katsman, Bicheng Gao, Serge Belongie

(ICML 2019 Workshop, Security and Privacy of Machine Learning) Adversarial Example Decomposition [Arxiv]

Horace He, Aaron Lou*, Qingxuan Jiang*, Isay Katsman*, Serge Belongie, Ser-Nam Lim

*indicates equal contribution



Quantitative Research Intern
May 2021 - Aug 2021

Conducted stochastic calculus research to develop improved implied volatility models for options.

Cornell Goldfeld Lab

Research Intern
May 2020 - Aug 2020

Worked with Prof. Ziv Goldfeld to develop a provable randomized smoothing method in the context of adversarial example defense research.


Research Intern
May 2019 - Aug 2019

Worked with Prof. Kristen Grauman.


Software Engineer Intern
May 2018 - Aug 2018

Intern on Facebook AI Research (FAIR) team.


Software Engineering Intern
May 2017 - Aug 2017

Worked on building RNNs for large-scale venue data extraction.


Software Engineering Intern
May 2016 - Aug 2016, May 2015 - Aug 2015

Worked on building Siamese and Triplet networks for appearance-based similarity learning.


Cornell University Artificial Intelligence (CUAI)

Research Advisor from June 2020 - present
Co-President from Aug 2018 - May 2020
Official Club Website
Press Release

Founded a club focused on undergraduate research and education in the fields of artificial intelligence and machine learning. The club is funded by and works directly with Facebook AI.

Service and Awards


  • ML conference reviewer for: NeurIPS 2021 (top 8% reviewer award), ICML 2021, NeurIPS 2020, ICML 2020 (outstanding reviewer award)
  • CV conference reviewer for: CVPR 2021 (outstanding reviewer award), CVPR 2020, ACCV 2020, CVPR 2019, ICCV 2019


  • NSF GRFP 2020 Honorable Mention
  • 2020 and 2019 CRA Award for Outstanding Undergraduate Researchers - Honorable Mention