Paddy is a versatile, biologically inspired evolutionary algorithm designed to efficiently optimize complex chemical systems while avoiding local minima.
- Robust and adaptable: Demonstrates strong performance across diverse benchmarks, from mathematical functions to chemical optimization tasks.
- Exploratory sampling: Effectively avoids early convergence by bypassing local optima, ensuring global solutions are identified.
- Chemical applications: Enables hyperparameter tuning, molecular generation, and experimental design, accelerating automated experimentation and discovery.
Paper · Code