Tue, Nov 28, 2017
Semi-supervised learning is a class of supervised learning tasks and techniques that also make use of unlabeled data for training – typically a small amount of labeled data with a large amount of unlabeled data.
The Data - STL-10 In this problem we are given 5,000 labeled images and 100,000 unlabeled images. - Colored 96x96 images - Labeled images belong to one of the 10 classes - The unlabeled 100,000 images contain other classes of images as well as the 10 given classes.
Sat, Nov 18, 2017
1. Problem Description We are given a rating matrix $R$ where only a small fraction of the entries are provided; otherwise the rest of them are missing. The task of a Recommender System is to predict those missing entries. As in most Machine Learning problems the assumption here is that there’s an underlying pattern of how users rate movies.
By the nature of the problem, $R$ is a sparse matrix, where the sparsity comes not from zero entries but from empty records.