Course syllabus

LT2203/LT2213/LT2813 VT 2025 Computational Semantics, Komputationell semantik, 7.5 credits, part of Master's Programme in Language Technology (MLT),  Applied Data Science Programme (N2ADS) and as a free-standing course.

Summary

In this course we will study and discuss different ways of representing meaning of words, sentences and conversations with computational methods, representations and models, using both top-down rule-based and data-driven techniques using machine learning. We will contrast and compare them with each other in terms of their strengths and weakness to represent meaning of natural language and the way they can be computationally modelled. We will examine how we can use them to draw inferences or reason with them computationally, and how they are suited for and can be used in different language technology tasks and applications.

Some topics that we will discuss are:

• rule-based representations in (formal) model theoretical semantics for natural language (as developed for example in Montague semantics);
• distributional representations in vector space models estimated from word distributions in texts;
• distributed representations such as word embeddings, language models and contextual word embeddings;
• semantic similarity, linguistic underspecification, compositionality, and reasoning in natural language;
• and others.

Course syllabi: (i) H2MLT and N2ADS: en , sv ; (ii) free-standing course: en, sv; (iii) PhD course: en.

Teachers

In the order of appearance... You can contact the lecturers by following the information on their webpages or you can send them a message through Canvas.

• Simon Dobnik (SD) (course organiser)
• Bill Noble (BN)
• Mattias Appelgren (MA)
• Ricardo Muñoz Sánchez (RMS)

Practical information

• Prerequisites and requirements to pass this course
• If you are taking this course as a free-standing course online
• Madelaine Miller (education coordinator)
• Annika Herrström (education administrator)
• Robert Adesam (systems engineer, help with mlt server access)

Course resources

• Slides, readings, and materials can be found under Modules
• Online classroomfor lectures, classes and seminars
• A reading list for the course
• Course materials on mltgpu.flov.gu..se: /srv/data/computational-semantics
• The previous edition of this course and the summary of the evaluation survey
• Related coursefor doctoral students

Location and schedule

The course will start on 27 March 2025 in J415 and online. We will then meet (with a couple of exceptions) every Thursday (for seminars or assignment classes) and on every other Monday (for lectures) at 13:15. For

All events start with an academic quarter.