A Hybrid Ant Colony Genetic Optimisation Framework for Automated CNN Architecture and Hyperparameter Tuning in Image Classification

Key Points

• To develop a hybrid optimization framework combining ant colony optimization and genetic algorithms for tuning CNN architectures and hyperparameters.
• Utilized a hybrid approach integrating ant colony optimization and genetic algorithms.
• Applied the framework to optimize Convolutional Neural Network (CNN) architecture for better performance.
• Performance evaluated on various image classification tasks.
• Achieved significant improvements in classification accuracy compared to standard tuning methods.
• Enhanced optimization led to faster convergence rates in finding optimal hyperparameters.
• Demonstrated improved model performance across different image datasets.