Mnist-Machine-Learning-Basics

Various methods for image classification of handwritten numbers on the well known image dataset MNIST.

Visualization techniques of the data points as cloud using graph theory and networks.

Run the Main.py first, and then if you want to see some plots run Plots.py

Dependencies for Main.py: numpy , matplotlib , sklearn, collections, itertools.

Dependencies for Plots.py: networkx , matplotlib, scipy, graphviz, pygraphviz

For Plots.py: Load and run Main.py first (better in Spyder or another python-IDE) and then run Plots.py

Report is available for further information on the Machine Learning algorithms.

Deep Learning

Various DL methods using TensorFlow/Keras

Convolutional auto-encoders for denoising

Research on convolutional auto-envoders on Cifar10 dataset for noising/denoising, and sound denoising as well.

Recurrent LSTM for language modeling

Deep recurrent neural-networks on language generation, translation and adding two short integers.

For more information please read the Report_DL.pdf file.