Guessing solutions to mathematical problems and learning maths from examples | Vidéo | Carmin.tv

00:00:00 / 00:00:00

Guessing solutions to mathematical problems and learning maths from examples

By Amaury Hayat

Appears in collection : FLAIM: Formal Languages, AI and Mathematics

In this talk, we will see how a deep language model can learn to guess solutions to master level problems in mathematics with an impressive accuracy (higher than most humans) and without any built-in mathematical knowledge. We will illustrate this in several examples, and we will discuss a possible extension to some mathematical problems for which no solution is currently known.

Information about the video

Date of recording 04/11/2022
Date of publication 04/11/2022
Institution IHP
Licence CC BY-NC-ND
Language English
Audience Researchers, Graduate Students, Students
Director(s) Alexandre Duplessis, Marco Perez
Format MP4
Venue IHP - Hermite Amphitheater

Domain(s)

Artificial Intelligence

Bibliography

*Charton, F., Hayat, A., & Lample, G. / Learning advanced mathematical computations from examples. International Conference on Learning Representations, September 2020. *Charton, F., Hayat, A., McQuade, S. T., Merrill, N. J., & Piccoli, B. / A deep language model to predict metabolic network equilibria. [arXiv:2112.03588] (https://doi.org/10.48550/arXiv.2112.03588).

MSC codes

Last related questions on MathOverflow

You have to connect your Carmin.tv account with mathoverflow to add question

Ask a question on MathOverflow

Copyright Carmin.tv 2026

Give feedback