Machine Learning for Natural Language Processing

Objective

This course aims to provide an overview of the techniques and applications of automatic natural language processing. It presents the evolution of the field, from its origins based on computational linguistics to the recent development of the major language models of generative AI.

Planning

Lecture 1 : Introduction

• Is NLP a solved problem ?
• Overview of computational linguistics
• Are LLMs NLP killers ?

Lecture 2 : Vector representations of words

• Bag-of-words and TF-IDF
• Latent Semantic Indexing
• Glove
• Word2Vec
• Document embeddings from word embeddings

Lecture 3 : Language models

• Statistical language models
• Neural language models
• Recurrent neural networks and LSTM
• Encoder-decoder
• Contextual word embeddings (ELMo)

Lecture 4:

• Transformers

Lecture 5 :

• LLM and applications

Lecture 6 :