Natural Language Processing

Module aims

In this module you will have the opportunity to:
- Learn about the foundations, building blocks and applications of Natural Language Processing (NLP), with an emphasis on approaches based on deep learning.
- Study the models used to represent words and word meanings.
- Use these representations to study classification tasks (e.g. sentiment analysis) and tagging tasks (e.g. part of speech tagging).
- View languages as sequences of variable length, from pure language models to machine translation models.
- See approaches that are based on modern neural machine learning algorithms, where linguistic information is provided by instances of uses of language.

Learning outcomes

Upon successful completion of this module you will be able to:
- describe and critically appraise NLP approaches used to support deep learning activities
- model words and languages using appropriate NLP representations
- discuss approaches to classification tasks (e.g. sentiment analysis) and tagging tasks (e.g. part of speech tagging).
- apply state-of-the art tools and techniques to solve real-world NLP problems

Module syllabus

Introduction to NLP  
Word meaning and representations
Classification tasks: spam detection, sentiment analysis
Language models (n-gram based, RNN, GPT)
POS tagging and language model
Parsing
Sequence to sequence models
Machine translation
Guest lecture(s) on advanced NLP topic(s)

Recommended modules: Introduction to Machine Learning; Python Programming.

Teaching methods

The material will mainly be presented in lectures, as well as in the weekly lab sessions, with the latter mainly designed to help you with the coursework. Where possible guest lectures will be used to provide you with additional viewpoints and introductions to advanced topics.

An online service will be used as an open discussion forum for the module.

Assessments

There will be one coursework that contributes 30% of the mark for the module. A typical coursework task will be a classification task where the input is language. This might involve building models to detect offensive language, patronising language, emotion, etc. There will be a final written exam, which counts for the remaining 70% of the marks.    

You will be provided with detailed written feedback on the coursework.

Module leaders

Dr Marek Rei
Mr Joe Stacey
Mr Nihir Vedd
Mr Nuri Cingillioglu

Reading list

To be advised - module reading list in Leganto