HarmanDotpy/ML-Assignment2

Implementations of svm's, smo algorithm, neural networks from scratch in python, standard neural network architectures like alexnet, lenet in pytorch, and attempts at zero-shot learning and domain adaptation(paper implementation), as part of the solutions for Assignment 2 of the course ELL409 : Machine Intelligence and Learning taught by Professor Prathosh AP

ML-Assignment2

Solutions for Assignment 2 of the course ELL409 : Machine Intelligence and Learning taught by Prof. Prathosh A.P.

Find the Final Report here: Complete Report

Find the Problem statements here: Problem Statements

Glimpses from the assignment

Fitting SVM's

(Old Data = Original Data with Support Vectors Marked in Yellow New Data = Support vectors removed from original data and SVM refitted, new support vectors marked in Yellow)

Old Data	New Data

MNIST Visualization

PCA	tSNE