Master 2024/2025

Deep Learning for Natural Language Processing

Type: Elective course (Math of Machine Learning)

When: 1 year, 4 module

Open to: students of one campus

Language: English

Course Syllabus

Abstract

The course is about modern models for natural language processing. The new generation of neural network-based methods based on deep learning has dramatically improved the performance of a wide range of natural language processing tasks, ranging from text classification to question answering. The course covers the basics and the details of successful models and methods for natural language processing based on neural networks, starting from the simple word embedding models, such as word2vec, all the way to more sophisticated language models, such BERT. Special attention is given to models based on the Transformer architecture, such as GPT, T5, BART, etc.