Language Models

Course goals

After completing the course students can name, illustrate and apply methods and techniques in the area of language and speech models. Students are able to apply the learned techniques through the modeling and evaluation of language applications in Python. Students are able to assess and comment on existing applications as well as applications they implemented themselves.

Content

This course used to be offered as KI3V24001 Computational Linguistics. Courses KI3V24001 and KI3V26003 are equivalent in terms of content; it is not allowed to take both courses as part of your curriculum.

Language models automate linguistic processes for a variety of AI-related applications. These applications have increasingly become an integrated part of our daily activities. During the lectures we will discuss several probabilistic methods and techniques that are needed to develop these applications. We will focus on themes such as words, classification, and speech technology, as well as more more recent developments such as Large Language Models (as used in GenAI assistants such as ChatGPT, Gemini, and Claude). In the computer lab sessions a number of these techniques will be implemented in Python. Course topics include: N-gram language models, part-of-speech tagging, hidden Markov modeling, Bayesian classification, speech synthesis, automatic speech recognition, Large Language Models. 

 

Additional information

Please note: the time slot shown here is not yet final and may still be modified until the 3rd Wednesday in September.