TLDRocket
Sign in

Symbolic Regression

2 summarised stories about Symbolic Regression, each linking back to the original source. Browse all topics →

Thursday, 16 July 2026

Discovering Ordinary Differential Equations with LLM-Based Qualitative and Quantitative Evaluation

arXiv cs.AI 18 hours ago

Researchers developed DoLQ, a method that uses large language models to discover ordinary differential equations from observational data by combining qualitative and quantitative evaluation through a multi-agent architecture. The system achieved higher success rates than existing methods on multi-dimensional ordinary differential equation benchmarks. By incorporating domain knowledge assessment through LLMs alongside traditional quantitative metrics, the approach improves accuracy in recovering correct symbolic terms of equations.

Automatic Ordinary Differential Equations Discovery For Biological Systems Using Large Language Model Powered Agentic System

arXiv cs.AI 18 hours ago

Researchers developed MEDA, an AI system combining large language models and symbolic regression to automatically discover mathematical equations describing biological systems. The system successfully recovered correct state variables and achieved strong structural recovery across retrieval, extrapolation, and open-ended discovery tasks, with knowledge-guided constraints proving essential for producing biologically plausible models. This approach moves beyond data-fitting toward generating mechanistic models that can incorporate domain knowledge and automate parts of the scientific discovery process for biological systems.