Learning Advanced Mathematical Computations from Examples
Abstract
Using transformers over large generated datasets, we train models to learn mathematical properties of differential systems, such as local stability, behavior at infinity and controllability. We achieve near perfect prediction of qualitative characteristics, and good approximations of numerical features of the system. This demonstrates that neural networks can learn to perform complex computations, grounded in advanced theory, from examples, without built-in mathematical knowledge.
Cite
Text
Charton et al. "Learning Advanced Mathematical Computations from Examples." International Conference on Learning Representations, 2021.Markdown
[Charton et al. "Learning Advanced Mathematical Computations from Examples." International Conference on Learning Representations, 2021.](https://mlanthology.org/iclr/2021/charton2021iclr-learning/)BibTeX
@inproceedings{charton2021iclr-learning,
title = {{Learning Advanced Mathematical Computations from Examples}},
author = {Charton, Francois and Hayat, Amaury and Lample, Guillaume},
booktitle = {International Conference on Learning Representations},
year = {2021},
url = {https://mlanthology.org/iclr/2021/charton2021iclr-learning/}
}