Admission requirements
Not applicable; desired background includes exposure to Python, machine learning and NLP (e.g. through the BA course Natural Language Processing (LIACS), and electives in the MA track Computational Linguistics).
Description
In this course, students learn about the theory and practice of a variety of machine learning algorithms for natural language processing (NLP). The course starts with a 4-class block discussing the general field, mainstream algorithms (including SVM, Gradient Boosted Trees, memory-based learning) and evaluation measures, before zooming in (the next 8 classes) on deep learning-based NLP. In this course, the leading application will be Conversational AI.
Course objectives
This course brings master students up to speed with current machine learning-based approaches to NLP, and introduces them to the application field of Conversational AI (chatbots, conversational agents). Students engage in practice by making (non-graded) weekly assignments using readily available machine learning toolkits, and discussing their results on shared data in class. The focus of the course is on application; students are not expected to do heavy programming. Occasional programming for data conversion, slight adaptation of existing tools and analysis may be necessary though (for which adequate support will be arranged by the teacher).
Timetable
The timetables are available through My Timetable.
Mode of instruction
Seminars and in-class discussion of practical assignments.
Assessment method
The assessment method addresses:
A group paper (describing an experiment)
A presentation of the group paper, in which individual contributions are presented by the students.
Weighing
The final score consists of a weighted average of
1. the group paper (scored for the group -max 4 persons- as a whole),
2. the personal presentation of the individual contribution to the paper, through a (group)
Powerpoint,
3. the individual contribution to the paper (which should be identifiable, like a section)
The personal part of the paper presentation is weighted as 1, the paper itself is weighted as 2, and the personal contribution to the paper is weighted as 3 (the total score divided by 6 gives the final grade).
Resit
A resit consists of rewriting the personal contribution to the paper, and a presentation of that contribution.
Reading list
Readining consists of papers to be administered through Brightspace and (optional)
Stephan Raaijmakers, Deep learning for Natural Language Processing. Manning, 2022. Epub and (later in 2022) hardcopy.
Registration
Enrolment through MyStudyMap is mandatory.
General information about course and exam enrolment is available on the website.
Contact
For substantive questions, contact the lecturer listed in the right information bar.
For questions about enrolment, admission, etc, contact the Education Administration Office: Reuvensplaats