Course at University of Bielefeld, held in Summer Semester 2022. Lecturer: Steffen Eger
- 15.08.2022: Deadline extension, see Lernraum+
- 29.07.2022: When the term papers are due, submit them via Lernraum+. See the latest updates on Lernraum+ for more information.
- 02.06.2022: New zoom for today's lecture (only this time); see the LernRaum+
- 22.05.2022: tutorial on 30.05. will be a general QA session on the structure of term papers and presentations
- 22.05.2022: there is no tutorial on 6.06., as it's a holiday. The class on 02.06. will cover both lecture and tutorial elements as a replacement
- 22.05.2022: voting for presentation slots is open next week; see the moodle
- 04.05.2022: There's a new zoom link for the lecture; see the moodle
- 29.04.2022: See the moodle announcement concerning Lecture 5
- 25.04.2022: Voting for topics available on moodle; ends on Saturday 12:00.
- 25.04.2022: New "Sprechstunde"
- Mondays, 16:15-17:45 (let me know at least 1 hour in advance)
- Thursdays, 15:30-17:00 (only if you let me know by Monday night at the very latest)
- 23.04.2022: I will try to also stream the next in-class lecture, if possible
- 21.04.2022: Next Monday in the tutorial, we will present the topics from which you can choose. Try to be present, but if you cannot be present, check the slides.
- Vote for your preferred topics within (roughly) 5 days following Monday.
- 21.04.2022: Lecture slides (Lecture 3) are minorly updated - I typically update them after the lecture, please take a look
- 18.04.2022: Tutorial is moved to Mondays, 14-16. Next tutorial will take place on 25.04.2022
- 13.04.2022: Lernraum is now available
- 13.04.2022: Solutions for Tutorial 1 have been posted on Lernraum
- 13.04.2022: A survey for finding the date of the second part of the course (either lecture or tutorial) is linked on Lernraum. It's open until this Sunday.
- Lecture 1 (07.04.2022), "Introduction": Lecture 1 is in-class
- Lecture 2 (14.04.2022), "ML principles" takes place online
- Inverted classroom videos (watch before you come to class):
- Video 1: Basic Concepts of ML
- Video 2: Supervised and Unsupervised ML
- Video 3: K nearest neighbors
- Video 4: Overfitting
- Video 5: Regularization
- Corresponding IC slides
- Lecture 3 (21.04.2022), "Backpropagation" is in-class
- Lecture 4 (28.04.2022), "Word Embeddings Part 1" takes places online
- IC Video (watch before you come to class):
- Word2Vec, Stanford
- Corresponding IC slides
- Lecture 5 (05.05.2022), "Dependency Parsing" takes place online
- IC Video (watch before you come to class):
- Dependency Parsing, Stanford
- Corresponding IC slides
- Lecture 6 (12.05.2022), "Word Embeddings Part 2" takes place online
- Optional background video: BERT
- Lecture 7 (19.05.2022), "Word Embeddings Part 3" is in-class (hybrid streaming, if possible)
- Lecture 8 (02.06.2022), "Convolutional Neural Networks" is in-class (hybrid streaming, if possible)
- Note: for hybrid streaming we will use the zoom link of the tutorial
- Lecture 9 (09.06.2022), "Recurrent Neural Networks" is in-class (hybrid streaming, if possible)
- Lecture 10 (23.06.2022), "Sequence-to-Sequence Models, Text Generation, Evaluation Metrics" is online: part1, part2
- Background videos:
- Lecture 11 (04.07.2022), "Efficiency, Explainability, Adversarial Attacks" is online
- Lecture 12 (14.07.2022), "Guest Lecture" is online
- Tutorial 1 (12.04.2022), 8:15-9:45
- zoom link (temporary): zoom
- Maybe we will move the tutorial to another slot on Tuesdays; I will send around a doodle as soon as possible
- Exercise Sheet
- Tutorial 2 (25.04.2022), 14:15-15:45
- zoom link (temporary): zoom
- Topic selection
- Exercise Sheet
- Template code
- Tutorial 3 (2.05.2022), 14:00-15:30
- zoom link (temporary): zoom
- Exercise Sheet
- Template code
- Tutorial 4 (9.05.2022), 14:15-15:45
- zoom link (recurring): zoom
- Exercise Sheet
- Template code
- Tutorial 5 (16.05.2022), 14:15-15:45
- zoom link (recurring): zoom
- Exercise Sheet
- Template code
- Tutorial 6 (23.05.2022), 14:15-15:45
- zoom link (recurring): zoom
- Exercise Sheet
- Template code (updated)
- Tutorial 7 (2.06.2022), 13:15-13:45
- zoom link (recurring): zoom
- Exercise Sheet
- Tutorial 8 (13.06.2022), 14:15-15:45
- zoom link (recurring): zoom
- Exercise Sheet
- Template code
- In-class lectures take place in CITEC Raumnummer: 0.007
Tutorial planned for Tuesdays 8:15 - 9:45, now Mondays 14-16- Tutorials are on zoom
- Lecture is hybrid, i.e., some lectures will take place live in Bielefeld, while the others will take place via zoom
- More information: see Lecture 1
- To pass the course, you need to work on an NLP problem (with deep learning), write a term-paper (8-10 pages) and present
- You can present on Mondays or on Thursdays
- Group work is in principle allowed
- Term paper: Mandatory, you must use this template
- Presentations: Optional, you may use this template or another one of your own choice
See e.g. https://github.com/SteffenEger/dldh/ (click through the years for the "Best term papers" in each year)