This repository is created for sharing study notes for the nlp class cs224n
【content】
prerequisite:
- linear algebra review
- probability theorem review
- convex optimization review
week 1: word2vec, skip-gram
- lecture 1: class introduction
- lecture 2: word2vec
- word2vec tutorial: skip gram neural network architecture for Word2Vec
- lecture notes01: Natural Language Processing. Word Vectors. Singular Value Decomposition. Skip-gram. Continuous Bag of Words (CBOW). Negative Sampling. Hierarchical Softmax. Word2Vec
- paper: Distributed Representations ofWords and Phrases and their Compositionality: subsampling, negative sampling, method for finding phrases
- paper: Efficient Estimation of Word Representations in Vector Space: dealing with large data set, achieving large improvements in accuracy at much lower computational cost
week 2: GloVe, word vector evaluation
- lecture 3: finish word2vec, GloVe intro
- lecture notes02: GloVe, evaluation, Window classification
- paper: GloVe_Global Vectors for Word Representation
- paper: Evaluation methods for unsupervised word embeddings
- paper: Improving Distributional Similarity with Lessons Learned from Word Embeddings
week 3: classification, backward propagation
- lecture 4: classification
- lecture 5: multi-layer NN, backward propagation
- lecture notes03: Neural networks. Forward computation. Backward propagation.
- gradient notes: Computing Neural Network Gradients
- paper: review-differential-calculus
- paper: Natural Language Processing (almost) from Scratch
- paper: backprop_old