← Back to Projects

🧬 ANN Trained using Genetic Algorithm

Summary:
Implemented an Artificial Neural Network whose weights are optimized using a Genetic Algorithm. The project explores alternative training approaches to improve convergence and to escape local minima typical in gradient-based methods.

My Role: Algorithm developer and experimenter

Technologies Used: C++, numerical optimization, evolutionary strategies

GitHub Repo: View Project

Key Achievements / Outcome

• Implemented chromosome encoding for neural weights and mutation/crossover strategies.
• Demonstrated improved robustness on synthetic benchmark datasets.