This repository contains a link to my thesis titled Word Vector Representations, Embedding Models and Applications, along with some implementations:
- a sample Vector Space Model pipeline (building a word--word co-occurrence matrix using the Reuters Corpus, performing truncated SVD on it, plotting word vectors and computing cosine similarity)
- word2vec implementation from scratch, based on Assignment2 of CS224n at Stanford University
- an example of an extrinsic task -- a feed-forward neural net to classify movie reviews from the Stanford Sentiment Treebank dataset