Multi-view Symbolic Regression: How to learn laws from examples
par
Amphi Recherche
Symbolic Regression is a data-driven method that searches the space of mathematical equations with the goal of finding the best analytical representation of a given dataset. However equations built with traditional Symbolic Regression approaches are limited to describing one event at a time, and thus does not constitute general laws but rather individual event descriptions.
I will present an adaptation of Symbolic Regression that is capable of recovering a common parametric equation hidden behind multiple datasets generated using different parameter values. We call this approach Multi-view Symbolic Regression (MvSR). I will demonstrate its efficiency on a challenging artificial benchmark and further highlight its potential by applying it on a variety of real scientific datasets, namely astrophysics, chemistry and finance. The resulting parametric equations are able to correctly describe the examples from which they were built as well as other unseen similar examples.
Reference: Russeil et al., 2024, Multi-View Symbolic Regression, arXiv:cs/2402.04298
