MTTTS16 Learning from Multiple Sources, Fall 2019 periods I and II
Course Contents
Also available in the Curriculum Guide for the course.
Contents. Learning from multiple sources denotes the problem of jointly learning from a set of (partially) related learning problems, views, or tasks. This general concept underlies several topics of research, which differ in terms of the assumptions made about the dependency structure between learning problems. During the course, we will cover a number of different learning tasks for integrating multiple sources and go through recent advances in the field. Examples of topics covered by the course include data fusion, transfer learning, multitask learning, multiview learning, and learning under covariate shift.
Learning Outcomes. After the course the student is familiar with several settings related to learning from multiple sources, and is familiar with a selection of important approaches and methods used for learning in each setting.
Passing the Course
To pass the course, you must pass the exam and complete a sufficient number of exercises from the exercise packs. Exercise packs will be released during the course.
Preliminary grading scheme (note: preliminary information only, may change!): the exercise packs are graded in total either as 0 (fail) or as a fractional number between 1 and 5 (such as 1.34). The exam is similarly graded either as 0 (fail) or as a fractional number between 1 and 5. The total grade of the course is computed as round(0.8*ExamGrade + 0.2*ExercisesGrade), so that e.g. 4.51 rounds up to 5 and 4.49 rounds down to 4.
Teaching Schedule
For general schedule information, please see the Curriculum information for the course for the Academic Year 2019–2020 (choose the “Completion Options” tab and expand the different components).
Preliminary Lecture Schedule
The regular lecture time is Tuesdays 14-16 in room Pinni B0016. Some changes in lecture dates may be possible – if changes are made they will be announced here and by email.
| 3.9. | L1: Introduction to the course, preliminaries. Material: lecture slides. Lecture video (audio only) available on the course Moodle page. |
| 10.9. | (no lecture, university opening ceremony) |
| 17.9. | L2: Basic Canonical Correlation Analysis. Material: lecture slides. Lecture video available on the course Moodle page. |
| 24.9. | L3: Basic multitask learning with neural network arrangements. Material: lecture slides. Lecture video available on the course Moodle page. |
| 1.19. | L4: Transfer learning. Material: lecture slides. Lecture video available on the course Moodle page. |
| 8.10. | L5: Transfer learning, continued, and probabilistic CCA. Material: lecture slides. Lecture video available on the course Moodle page. |
| 16.10. at 10:15-12, room B3108 (note updated day+time+room) | L6: Kernel CCA and other variants. Material: lecture slides. Lecture video available on the course Moodle page. |
| 22.10. | L7: Multitask learning with task clustering or gating. Material: lecture slides. Lecture video available on the course Moodle page. |
| 29.10. | continuation of L7. Material: same slides as the previous lecture. Lecture video available on the course Moodle page. |
| 5.11. | L8: Multitask learning with kernel methods and nonparametric models. Material: lecture slides (Updated Nov 19). Lecture video available on the course Moodle page. |
| 12.11. at 10:15-12, room B1083 | L9: Multi-view learning for classification by Co-training. Material: lecture slides. Lecture video available on the course Moodle page. |
| 19.11. | L10: Co-training continued; no new slides. Lecture video available on the course Moodle page. |
| Wednesday 27.11. at 12:15-14, room B0016 (note updated day+time) | L11: co-training concluded: disagreeing views; semisupervised multi-task learning; self-taught learning. Material: lecture slides. Lecture video available on the course Moodle page. |
| 3.12. | L12: Domain adaptation. Material: lecture slides. Lecture video available on the course Moodle page. |
| 10.12. | L13: Domain adaptation concluded + learning sample correspondence. Material: lecture slides. Lecture video available on the course Moodle page. |
| 17.12. at 14-18 in room B3107 | First exam. |
Informal Home Assignments
Some assignments to help you think about the lecture material will be published here. Return solutions by email to the lecturer.
Exercise Packs
Exercise packs released during the course will appear here.
- Exercise pack 1 has been released and is available on the Moodle page. Return deadline: November 22.
- Exercise pack 2 has been released and is available on the Moodle page. Return deadline: January 10.