Natural Language Processing

Course Objective

Natural language processing (NLP) is a highly dynamic research field
that aims at analyzing natural language using computational models and
methods. This course provides an introduction to the principles of the
field and has two main objectives:

1) Students learn the fundamentals for developing a system for a common
NLP task.
- They can analyze and interpret the linguistic information provided by
common pre-processing pipelines and are familiar with the core NLP
terminology. (Knowledge & Understanding)
- They can describe a common NLP classification setup and evaluate the
results. (Making Judgements)
- They can explain the functionality of neural networks and their use in
NLP. (Communication)
- They can explain the role of pre-trained language models in NLP and
apply them to a task. (Applying knowledge and understanding)

2) Students learn to systematically analyze and interpret NLP models for
a specific task.
- They can describe the linguistic challenges associated with the task.
(Making judgments)
- They can describe how the task is approached by a state-of-the-art
model. (Knowledge and understanding)
- They are able to understand, evaluate, and interpret the output of a
model. (Applying knowledge and understanding)
- They can compare the output of different models and analyze their
strengths and weaknesses. (Lifelong learning skills)

Course Content

Natural Language Processing operates on the interface between

linguistics and computer science. In order to get computers to deal well

with natural language, it is important to understand both how language

works and how computational methods work. Computational linguists work

on this interface and have developed methods and technologies for

language analysis.

This course covers technologies and computational models for core

domains of natural language processing. Students are trained to find,

process, and understand the latest developments in this rapidly

advancing field. The course includes practical components that require

students to work with Python code to explore NLP models. They learn to

apply them to new data and reflect on the effect of modeling decisions

on the performance.

Additional Information Teaching Methods

The course consists of two sessions per week. We generally alternate
between more theoretical lectures and interactive practical classes in
smaller groups.

Method of Assessment

The course is evaluated by practical assignments (50%) and an exam
(50%). Both components need to be graded at least with a 5.5. Retakes of
practical assignments to improve the grade are not allowed. Practical
assignments might be organized as a mix of group projects and individual
contributions.

Literature

This information will be specified on Canvas.

Additional Information Target Audience

Master Artificial Intelligence, also suitable for Master Business
Analytics and other Master students in Computer Science

Recommended background knowledge

Students must have acquired programming skills in Python. For students

with no background knowledge of machine learning, additional reading

might be required as the fundamental concepts will only be briefly

introduced in the course.