Computer Modelling of Language

Aim and Content

The content of the course may vary, but will treat theories and methods for the development and use of different types of language models. Relevant topics are, for instance, knowledge representation, semantic networks, pattern recognition, goal-driven and data-driven planning, advanced search strategies, heuristics, rule based systems, frames and inheritance hierarchies, constraint propagation, logic and resolution, machine learning, neural networks, probabilistic and statistical models or evolutionary models.

This course unit may also be appropriate for students in computer science, information science and humanities computing.

The goal of the course is to make students familiar with modeling methods, for instance, from artificial intelligence and machine learning.

Learning Outcomes

After successfully completing this course, students will be able to understand the relevance of the treated methods for modeling language processes in computational linguistics, psycholinguistics and language technology. Students will be able to apply modeling approaches to scientific and technological problems in natural language processing. Students will be able to program and test models based on these methods.

Course offered (semester)

Autumn

Language of Instruction

English if students not speaking Norwegian attend, otherwise Norwegian.

Course Unit Level

Master.

Department

Department of Linguistic, Literary and Aesthetic studies.

Access to the Course Unit

Open to all who have been admitted to the master's program in Natural Language Processing at the University of Bergen.

Pre-requirements

None.

Recommended previous knowledge

Some computer programming experience.

Teaching Methods

Interactive lectures with computer exercises, about 3 hours per week. Supervision of the term paper project.

If fewer than five students are registered to a course, the department might reduce the teaching, please see the department¿s guidelines regarding this on ¿My Space¿. For courses where this is a possibility, the students get information about this at the beginning of the semester, and before the deadline regarding semester registration February 1 / September 1.

Compulsory Requirements

Supervision in connection with the term paper is required, and the student is required to have met for supervision at least once, where the supervisor has been able to read a draft of the paper in advance.

Assessment methods

Supervised term paper in the form of a programming assignment and four-hour written exam. Both assessment forms are weighed equally in order to determine the final grade. The term paper must be handed in before the student can take the written exam.

Grading Scale

A-F.

Reading List

Reading comprises about 400 pages, mainly consisting of an introduction to artificial intelligence and a scientific description of methods used within the field, with emphasis on cognitive modeling. Computer tools are used for the exercises.

Place of Teaching

Bergen.

Course Unit Evaluation

Evaluation of the course will be carried out in accordance with the University of Bergen's quality assessment system.

Contact Information

advice@lle.uib.no