Physical Symbolic Optimization

Wassim Tenachi, Rodrigo Ibata, Foivos I. Diakogiannis
12/6/2023
cs.LGastro-ph.IMcs.SC

No official code found. Best implementation: WassimTenachi/PhySO (1,975 stars, MIT, last commit 2026-02-21).

Abstract

We present a framework for constraining the automatic sequential generation of equations to obey the rules of dimensional analysis by construction. Combining this approach with reinforcement learning, we built $Φ$-SO, a Physical Symbolic Optimization method for recovering analytical functions from physical data leveraging units constraints. Our symbolic regression algorithm achieves state-of-the-art results in contexts in which variables and constants have known physical units, outperforming all other methods on SRBench's Feynman benchmark in the presence of noise (exceeding 0.1%) and showing resilience even in the presence of significant (10%) levels of noise.

View on arXivView PDF

Code Implementations(1)

WassimTenachi/PhySOCommunity100%
MIT

Physical Symbolic Optimization

1,975264Shell, PythonJan 27, 20238 months agoMIT
deep-learningequation-discoverymachine-learningphysicspython+2 more

Similar Papers

Deep symbolic regression for physics guided by units constraints: toward the automated discovery of physical laws

88% match1 repo2.0k2023

Class Symbolic Regression: Gotta Fit 'Em All

84% match1 repo2.0k2023

Deep Generative Symbolic Regression

83% match2 repos312024

StruSR: Structure-Aware Symbolic Regression with Physics-Informed Taylor Guidance

83% match2025

Neural Structure Embedding for Symbolic Regression via Continuous Structure Search and Coefficient Optimization

83% match2026

Vertical Symbolic Regression via Deep Policy Gradient

82% match1 repo2024

Cite this paper

@article{tenachi2023physical,
  title  = {Physical Symbolic Optimization},
  author = {Wassim Tenachi and Rodrigo Ibata and Foivos I. Diakogiannis},
  year   = {2023},
  eprint = {2312.03612},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2312.03612v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2312.03612.svg)](https://codeofpaper.com/paper/2312.03612)

Discussion