Skip to content

Repository files navigation

MCP-Driven LLM Agent for Autonomous Optimization of Physics-Informed Symbolic Regression

Overview

An autonomous LLM agent built using the Model Context Protocol (MCP) that optimizes symbolic regression on real industrial datasets from particle technology research. This project was done as Master's Thesis in Institute of Particle Technology within TU Braunschweig

The Problem

Traditional symbolic regression requires manual parameter tuning — a time-consuming trial-and-error process with no guarantee of physical consistency.

The Solution

An agentic loop where the LLM:

  • Evaluates regression outputs against physics constraints
  • Detects overfitting and unphysical terms
  • Autonomously adjusts PySR parameters & search grammars
  • Iterates until convergence on a physically meaningful equation

Architecture

image

Tech Stack

Python · PySR · MCP · Gemini 2.5 Flash · PyTorch · Scikit-learn

Key Results

  • Successfully derived physically consistent equations from 6 industrial targets
  • Reduced manual iteration cycles significantly
  • Composite physics scoring across 7 evaluation dimensions

About

This project was done as Master's Thesis in Institute of Particle Technology within TU Braunschweig. An autonomous LLM agent built using the Model Context Protocol (MCP) that optimizes symbolic regression on real industrial datasets from particle technology research.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages