Skip to content

Implementations of optimization techniques, sampling methods and evolutionary algorithms

Notifications You must be signed in to change notification settings

iremddemir/computational_intelligence

Repository files navigation

Computational Intelligence

Overview

This repository showcases the coursework from the Computational Intelligence course completed at Vrije University during Spring 2023.

Projects

1. Optimization Techniques

  • Exploration of gradient-based and derivative-free optimization methods. The project focuses on minimizing a complex function, providing insights into the nuances of optimization techniques in computational intelligence.

2. Sampling Methods

  • Implementation of Metropolis-Hastings and Simulated Annealing algorithms. The assignment offers a comparative study of these methods applied to a mixture of Gaussians, showcasing their behaviors and differences.

3. Evolutionary Algorithms

  • Application of evolutionary algorithms to a non-differentiable optimization task. This project includes the development and analysis of recombination and mutation operators, along with selection mechanisms, focusing on optimizing a black-box function.

4. Neural Networks and Image Classification

  • Implementation of fully-connected and convolutional neural networks using PyTorch for image classification. This project emphasizes the understanding of neural network architectures, optimizers, and loss functions in deep learning.