Interpretable Relational Inference with LLM-Guided Symbolic Dynamics Modeling

TL;DR

COSINE is a novel framework that combines symbolic regression and large language models to infer interpretable interaction structures and dynamics from observed data, improving mechanistic understanding.

Contribution

It introduces a differentiable method that jointly discovers interaction graphs and symbolic dynamics, enhanced by LLM-guided hypothesis space adaptation.

Findings

Robustly recovers interaction structures from synthetic data.

Produces compact, mechanism-aligned dynamical expressions.

Demonstrates effectiveness on real-world epidemic data.

Abstract

Inferring latent interaction structures from observed dynamics is a fundamental inverse problem in many-body interacting systems. Most neural approaches rely on black-box surrogates over trainable graphs, achieving accuracy at the expense of mechanistic interpretability. Symbolic regression offers explicit dynamical equations and stronger inductive biases, but typically assumes known topology and a fixed function library. We propose \textbf{COSINE} (\textbf{C}o-\textbf{O}ptimization of \textbf{S}ymbolic \textbf{I}nteractions and \textbf{N}etwork \textbf{E}dges), a differentiable framework that jointly discovers interaction graphs and sparse symbolic dynamics. To overcome the limitations of fixed symbolic libraries, COSINE further incorporates an outer-loop large language model that adaptively prunes and expands the hypothesis space using feedback from the inner optimization loop. Experiments on synthetic systems and large-scale real-world epidemic data demonstrate robust structural recovery and compact, mechanism-aligned dynamical expressions. Code: https://anonymous.4open.science/r/COSINE-6D43.