Juan Pablo Vigneaux

AI and Mathematics

prof_pic.jpg

Since July 2025, I am a Research Associate in the Department of Linguistics at Northwestern University. I work with Jennifer Cole and Khalil Iskarous on mathematical models that combine prosody and meaning in connection with pragmatics.

I’m also working on interpretability of transformer-based LLMs, with a particular focus on the geometry of latent spaces and on compositionality. For further details, see the research section.

From October 2021 to December 2024, I served as an Olga Taussky and John Todd Instructor in Mathematics at Caltech, mentored by Matilde Marcolli. We co-organized the Information, Geometry, and Physics Seminar.

Previously, I was a postdoctoral researcher at the Max Planck Institute for Mathematics in the Sciences. In 2019, I obtained my Ph.D. in Mathematics at the University of Paris under the supervision of Daniel Bennequin.

Mathematically, I’ve worked on information theory in connection with category theory and geometry (metric geometry, geometric measure theory, …). Some notable interests, reflected in my publication record, are:

  • Characterization of information measures, mainly information topology.
  • Information dimension and measures with geometric structure.
  • Magnitude and diversity.

Here’s my academic CV (last update: February 2025).

latest posts

selected publications

  1. Journal
    Information theory with finite vector spaces
    Juan Pablo Vigneaux
    IEEE Transactions on Information Theory, 2019
  2. Journal
    Typicality for stratified measures
    Juan Pablo Vigneaux
    IEEE Transactions on Information Theory, 2023
  3. Journal
    Information structures and their cohomology
    Juan Pablo Vigneaux
    Theory and Applications of Categories, 2020
  4. Journal
    A formula for the categorical magnitude in terms of the Moore-Penrose pseudoinverse
    Stephanie Chen, and Juan Pablo Vigneaux
    Bulletin of the Belgian Mathematical Society - Simon Stevin, 2023
    To appear in Vol. 30, Issue 3