FluxDisco: Symbolic Regression for Stoichiometric Dynamical Systems via Monte Carlo Graph Search
Abstract
Dynamical symbolic regression methods identify governing differential equations from noisy data, balancing interpretability and predictive accuracy. However, standard methods often produce expressions that violate known physical laws. To address this, we propose FluxDisco, a physics-informed framework tailored for flux-based, stoichiometric ODE systems. By leveraging a known stoichiometry, we reduce the expression search space and ensure physical adherence. Our framework adapts the Monte Carlo Graph Search algorithm for the unique challenges associated with joint flux discovery of stoichiometric systems. We evaluate our method across a range of physical and biological systems, demonstrating its ability to accurately recover governing dynamics through interpretable equations.
Related
- New Symbolic Procedures in the Study of Dynamical Systems
- Learning Closure of Dynamical Systems with Kernel Ridge Regression
- Equation discovery for climate impact: emulating impact models for unexplored climate scenario with interpretable symbolic regression
- An Efficient Symbolic Algorithm for Computing Lyapunov Constants in Planar Switching Systems
- Symbolic Direct Construction of Augmented Systems for Levenberg-Marquardt Power Flow
BibTeX
@misc{durr2026fluxdisco,
title = {{FluxDisco: Symbolic Regression for Stoichiometric Dynamical Systems via Monte Carlo Graph Search}},
author = {Cassandra Durr and Alvaro Köhn-Luque and Chris Jewell and Lloyd A. C. Chapman},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.05207},
eprint = {2609.05207},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.05207},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - FluxDisco: Symbolic Regression for Stoichiometric Dynamical Systems via Monte Carlo Graph Search
AU - Cassandra Durr
AU - Alvaro Köhn-Luque
AU - Chris Jewell
AU - Lloyd A. C. Chapman
PY - 2026
DA - 2026/09/04
PB - arXiv
DO - 10.48550/arxiv.2609.05207
UR - https://arxiv.org/abs/2609.05207
AB - Dynamical symbolic regression methods identify governing differential equations from noisy data, balancing interpretability and predictive accuracy. However, standard methods often produce expressions that violate known physical laws. To address this, we propose FluxDisco, a physics-informed framework tailored for flux-based, stoichiometric ODE systems. By leveraging a known stoichiometry, we reduce the expression search space and ensure physical adherence. Our framework adapts the Monte Carlo Graph Search algorithm for the unique challenges associated with joint flux discovery of stoichiometric systems. We evaluate our method across a range of physical and biological systems, demonstrating its ability to accurately recover governing dynamics through interpretable equations.
ER -