🔬 Scientific Discovery Engine

The Scientific Discovery Engine is a powerful symbolic regression framework designed to rediscover fundamental physical laws from empirical data. Unlike standard machine learning, it prioritizes interpretability, physical consistency, and structural elegance.

🌟 Key Features

• Physics-Informed Discovery: Enforces strict dimensional consistency across Mass, Length, Time, and more. Penalizes "physics nonsense" to guide evolution toward plausible laws.
• Hybrid Optimization: Combines Genetic Programming (for structural discovery) with SciPy's curve_fit (for precise constant refinement).
• High Performance: Fully parallelized evaluation logic to utilize modern multi-core architectures.
• Dimensional Constants: Supports automatic discovery of unit-bearing constants (e.g., Gravitational constant, Gas constant).
• Automated Reporting: Generates publication-ready LaTeX reports covering the Pareto front of discovered formulas.

🚀 Quick Start

Installation

git clone https://github.com/trungminhdo4-glitch/algorithm-evolver.git
cd algorithm-evolver
pip install -r requirements.txt

Run the Demo

Experience the engine in action through the interactive CLI:

python demo.py

📊 Discovery Showcase

The engine has successfully rediscovered the following laws:

Experiment	Discovered Law	Physical Domain
Kepler's 3rd Law	$T^2 \propto a^3$	Orbital Mechanics
Ballistic Motion	$d = \frac{v^2 \sin(2\theta)}{g}$	Classical Mechanics
Ideal Gas Law	$PV = nRT$	Thermodynamics
Damped Oscillation	$x(t) = A e^{-\delta t} \cos(\omega t)$	Dynamic Systems

🏗 Architecture

The engine is built on a modular stack:

• core/: Evolutionary logic, physics kernels, and program simplification.
• problems/: Physical problem definitions and synthetic data generators.
• utils/: Fine-tuning modules and LaTeX exporters.

---

Developed for the intersection of Artificial Intelligence and Physical Sciences.