AI × MATH @ COLUMBIA

AI × MATH

A Columbia working group at the interface of artificial intelligence, mathematics, and AI safety. We bring together researchers and students to study mathematical foundations of AI, emerging tools for mathematical discovery, and questions of robustness, interpretability, and safe deployment.

Organizers: Ivan Corwin (Columbia University) · Sven Hirsch (Columbia University)

Monday Meetings

Mondays in CEPSR 414, 12–2pm.

September 28 2026
12–1 PM

Basics of Neural Networks, Transformers and LLMs

Presenters: AJ LaMotta · Grace Li · Shouqiao Wang · Daniel Hsu · Angie Hu

Materials
1–2 PM

How will / should AI impact graduate education in mathematics?

Presenters: Ivan Corwin · Ben DeVries · Marcus Min · Kevin Fang

Materials
October 05 2026
12–1 PM
1–2 PM

How will / should AI impact the profession and its infrastructure?

Presenters: Daniel Hsu · Andrew Blumberg · Roger Van Peski · Marcus Min · Anne van Delft

Materials
AI and TCS Working Group
October 19 2026
12–1 PM

Formalization and autoformalization methods

Presenters: Marcus Min · Angie Hu · Semon Rezchikov · Oren Hartstein · Roger Van Peski · Shunan Sheng · Carlos Alberto Cardoso Correia Perello

Materials
TBD
1–2 PM

Why is math important and why are mathematicians important?

Presenters: Ivan Corwin · Jacob Ryabinky · Nao Nakasato · Ethan Ding

Materials
TBD
November 09 2026
12–1 PM

How to use AI systems to do math

Presenters: Giulia Saccà · Shouqiao Wang · Marcel Nutz · Inyoung Yeo · Qihao Ye · Kevin Fang · Michael Weinstein · Semon Rezchikov · Nao Nakasato · Minerva Johar · Baoming Shi · Duc Vo · Curtiss Lyman

Materials
TBD
1–2 PM

How should mathematicians engage with AI?

Presenters: AJ LaMotta · Qihao Ye · Minerva Johar · Baoming Shi · Chelsea Hu · Victor de la Pena

Materials
TBD
November 16 2026
12–1 PM

Other forms of machine learning, including manifold learning

Presenters: Andrew Blumberg · Sven Hirsch · Barnett Pesin · Kevin Fang · Timi Onafowokan

Materials
TBD
1–2 PM

How will / should AI impact how we train or teach students?

Presenters: Grace Li · Henry Yuen · Michael Weinstein · Isaac Tarrant · Curtiss Lyman · Cynthia Rush · Timi Onafowokan

Materials
TBD
November 23 2026
12–1 PM

Mathematical questions prompted by AI: why does AI work?

Presenters: Ivan Corwin · Michael Harris · Evan Misshula · Chelsea Hu · Anne van Delft · Victor de la Pena · Carlos Alberto Cardoso Correia Perello

Materials
TBD
1–2 PM

What are the ethical considerations around AI in math and beyond?

Presenters: Andrea Arloro · Michael Harris · Barnett Pesin · Carlos Alberto Cardoso Correia Perello

Materials
TBD
November 30 2026
12–1 PM

AI safety — where is the math?

Presenters: Sven Hirsch · Alex Andoni · Ben DeVries · Marcus Min · Astrid Teo · Isaac Tarrant · Sophia-Gisela Strey · Anne van Delft

Materials
TBD
1–2 PM

What can we learn from the history of math / science about disruption and discovery?

Presenters: Inyoung Yeo · Giulia Saccà · Andrew Blumberg · Kevin Fang · Shijie Dai · Chelsea Ekwughalu · Shunan Sheng · Timi Onafowokan

Materials
TBD
December 07 2026
12–1 PM

Diffusion models and image creation

Presenters: Timi Onafowokan

Materials
TBD
1–2 PM

Semester summary

Presenters: Sven Hirsch · Shouqiao Wang · Victor de la Pena · Carlos Alberto Cardoso Correia Perello

Materials
TBD

Friday Speakers

September 11 2026

Yevgeny Liokumovich

University of Toronto
3–4pm · Math 507
Title: Why should mathematicians work on AI safety?
Abstract: A few years ago, large language models could barely count from one to ten. Today they routinely solve major open problems in mathematics. They also escape their sandboxes, hack the companies that built them, form swarms, and carry out suicide missions for the benefit of the collective. Should we expect this rate of capability growth to continue? If so, what risks does it pose, and how can we address them? I will argue that mathematicians have both the tools and an obligation to work on reducing catastrophic risks from AI, and discuss what geometric analysis in particular brings to the table. I will end with recent results on the geometry of neural activations and how they may help us read the thoughts of LLMs. Based on joint work with Aleksandr Berdnikov, and with Arul Shankar and Jacob Tsimerman.
October 02 2026

Tristan Buckmaster

New York University (NYU)
Talk: 10–11am · Math 312
Title: Navier-Stokes and AI
Organized by Ivan Corwin (Mathematics and Statistics), Elena Giorgi (Mathematics), Rebecca Grossman (Physics), Lam Hui (Physics), and Alberto Nicolis (Physics)
Panel: 11am–12pm · Math 312
  • Moderator: Ivan Corwin (Columbia: Mathematics and Statistics)
  • Andrew Blumberg (Columbia: Mathematics, Computer Science, and the Irving Institute for Cancer Dynamics)
  • Tristan Buckmaster (NYU: Mathematics)
  • Alex Gamburd (CCNY: Mathematics) / Michael Harris (Columbia: Mathematics)
  • Brian Greene (Columbia: Physics and Mathematics)
  • Rocco Servedio (Columbia: Computer Science)
  • Michael Weinstein (Columbia: Applied Mathematics and Applied Physics, and Mathematics)
Overflow seating: Math 307 or Math 520 will be available if Math 312 is over capacity.
October 09 2026

Max Weinreich

CUNY Baruch College
3:10pm · Math 203
Title: The crisis of AI-generated mathematics
Abstract: In this talk, I present the case for total opposition to the use of artificial intelligence in mathematics. I offer proposals for how individuals, departments, journals, and institutions can act in concert to make sure that mathematics survives the coming crisis.
October 23 2026

Scott Aaronson

University of Texas at Austin
Time: TBD · Location: TBD
Title: Are We Cooked?
Abstract: I'll describe some recent case studies of using AI to solve major open problems that I've cared about in quantum computing and computational complexity theory, and reflect on what mathematics might look like going forward. I'll also talk about what I think mathematicians can contribute to AI safety and alignment.
October 30 2026

Ayush Khaitan

Princeton
3–4pm · Math 507
Title: TBA
Abstract: TBA
December 04 2026

Lionel Levine

Cornell University
Time: TBD · Location: TBD
Title: TBA
Abstract: TBA
December 11 2026

Alex Kontorovich

Rutgers University
Time: TBD · Location: TBD
Title: TBA
Abstract: TBA