HIO; Specialization 2: Representation Learning for NLP (RL4NLP)

Representation Learning for Natural Language Processing (RL4NLP) introduces students to modern machine learning techniques for representing language units such as words, sentences, and documents in continuous vector spaces. The course covers both foundational and advanced embedding methods, including static and contextual representations, attention mechanisms, and language models, with a particular emphasis on unboxing and interpreting the internal representations and processing of large language models.

Through a combination of theory-driven lectures and hands-on project work, the course strengthens students’ critical thinking skills, scientific reasoning, and ability to document and communicate research results in a clear and rigorous manner. Collaborative assignments further foster teamwork and experience with research-oriented workflows.
 

The course is particularly relevant for students in the IT & Cognition program, computer science, and students from other disciplines with a sufficient mathematical and computer science background. It can also be seen as an initial step for students who intend to pursue a master’s thesis or research projects in natural language processing, machine learning, or related areas.