Capita Selecta of Subsymbolic AI

6 ECTS credits
150 h study time

Offer 1 with catalog number 4024145ENR for all students in the 1st and 2nd semester at a (E) Master - advanced level.

Semester

1st and 2nd semester

Enrollment based on exam contract

Impossible

Grading method

Grading (scale from 0 to 20)

Can retake in second session

Yes

Taught in

English

Faculty

Faculty of Sciences and Bioengineering Sciences

Department

Computer Science

Educational team

Lynn Houthuys (course titular)

Activities and contact hours

26 contact hours Lecture
26 contact hours Seminar, Exercises or Practicals
120 contact hours Independent or External Form of Study

Course Content

In this seminar the students get insights about recent developments in the field of subsymbolic AI, more specifically the field of kernel models. This is an advanced course about selected topics from the state of the art in kernel models and techniques, including multi-modality and deep learning.
The first part of the course will cover the principles of kernel methods, the support vector machine framework and kernel methods for various learning tasks.
The second part of the course will cover state of the art topics, such as multi-modality and deep learning, as well as aspects and discussions on fairness and explainability. The exact content of this part can vary each year, in order to reflect the latest research and practise. Students will have to study a certain topic in a small group, and present their findings to their fellow students.
During the practical sessions students will work on python exercises related to the material from the lectures. Additionally homework problems will be given which will count towards their final grade.

Additional info

None

Learning Outcomes

General competencies

Knowledge and Insight:
The student can explain the working of kernel methods in general, and the working of the covered kernel methods in specific.
The student shows insight in the recent trends regarding kernel-based learning.

Application of Knowledge and Insight:
Given a concrete learning problem, the student selects an appropriate kernel-based method, applies it correctly and is able to correctly analyse the performance.

Judgement Shaping:
The student collects and interprets literature about subsymbolic AI, and is able to understand it to a sufficient level in order to apply it correctly in appropriate problems.

Communication:
The student communicates about kernel-based problems and can report and present the results of their experiments, with both experts and non-experts.

Learning Skills:
After having attended this course, the student has the necessary knowledge to independently investigate a given research topic based on specific research papers and other resources.

Grading

The final grade is composed based on the following categories:
Other Exam determines 100% of the final mark.

Within the Other Exam category, the following assignments need to be completed:

Other exam with a relative weight of 1 which comprises 100% of the final mark.

Note: See 'Additional information regarding evaluation'

Additional info regarding evaluation

The final grade consist of three parts:

40%: Oral exam where students can use their notes
40%: Homework problems. At the end of the semester students will need to submit a written report about their solutions. During an oral examination they will be asked further questions about their specific work. Their grade is based on the report as well as the oral examination.
20%: Presentation of a selected topic, graded both on content and delivery.

Allowed unsatisfactory mark

The supplementary Teaching and Examination Regulations of your faculty stipulate whether an allowed unsatisfactory mark for this programme unit is permitted.

Academic context