SGN-41007 Pattern Recognition and Machine Learning
- [14.11.2019]The mandatory assignment (competition) pass requirements consist of 2 parts:
- Training a sklearn model with CNN feature extractor. Deadline Sunday 24.11.2019 at 23:55. Return to Moodle. See also video lecture of the assignment (25 min).
- Training a CNN model. Deadline Sunday 8.12.2019 at 23:55. Return to Moodle. See also a video lecture about the assignment.
- [8.11.2019] The course competition groups have now been formed. The first email is responsible for contacting the others and set up the first meeting.
- [5.11.2019] The course competition has started. Register yourself (with or without a group) here (deadline Friday, Nov. 8 at noon).
- [21.10.2019] Some additional material is available at the course github repo.
- [14.10.2019] Starting 28.10, the Monday lecture will move to Tuesdays, at 14-16, room TB104. The first lecture is still held on Monday, 21.10.
- [23.9.2019] First lecture will be on October 21st, 2019 at 12:15-14:00 in room TB109.
- [23.9.2019] There will be a video recording of each lecture (live + stored).
- [23.9.2019] The registration for weekly exercise groups is now open in POP and closes after the first lecture. Reserve your seat early enough.
- [23.9.2019] Common questions: Q: Can I return exercises by email? (A: Unfortunately not; the course is too big for that).
Q: Are exercises from last year still valid? (A: Yes they are).
There are two lectures every week (total 14 lectures: first time 21.10.2019, last time 5.12.2019):
- Tuesdays at 14-16 in TB104. First week on Monday 12-14 TB109.
- Thursdays at 12-14 in RG202
A video recording of each lecture will be provided below. Links to slides
and videos appear after each lecture. There is also a live broadcast, that should be
accessible via the Echo360 service. Recorded videos are available in the below list without a need to login.
- Monday 21.10.2019: Course organization, Introduction to Python. [video]
- Thursday 24.10.2019: Estimation Theory. [video]
- Tuesday 29.10.2019: Detection Theory. ROC and AUC. [video]
- Thursday 31.10.2019: Precision and Recall. Classification: The K-NN classifier and linear classifiers. [video]
- Tuesday 5.11.2019: Linear Classifiers: The LDA and the role of projection. [video]
- Thursday 7.11.2019: Linear Classifiers: SVM and the kernel trick; Logistic Regression. [video]
- Tuesday 12.11.2019: Neural networks. [video]
- Thursday 14.11.2019: Convolutional networks. [video]
- Tuesday 19.11.2019: Convolutional networks, Recurrent nets. [video]
- Thursday 21.11.2019: Recurrent networks. Applications of deep learning. [video]
- Tuesday 26.11.2019: Recurrent networks. Applications of deep learning. [video]
- Thursday 28.11.2019: Random Forest. Other ensemble methods in sklearn:
ExtraTreesClassifier, AdaBoostClassifier and
- Tuesday 3.12.2019: Performance assessment: Cross-validation.
Regularization, feature selection. [video]
- Thursday 5.12.2019: Visiting lectures from companies: Scandit Xiaomi and VTT. [video]
Exercise tasks will appear below, and sessions are once every week.
- Exercise 28.10. - 1.11.: Questions. Solve Python exercises in the exercise class. Solve pen and paper tasks beforehand.
- Exercise 4.11. - 8.11.: Questions.
- Exercise 11.11. - 15.11.: Questions.
- Exercise 18.11. - 22.11.: Questions.
- Exercise 25.11. - 29.11.: Questions. (updated on the evening of Nov 21st)
- Exercise 2.12. - 5.12.: Questions.
Exercise scores here.
Groups take place at the following times:
Registration for the groups is required. Exercises consist of theory and computer exercises. You can use the classroom computer or your own
laptop. Installation of necessary software is straightforward: Anaconda Python
distribution should contain all necessary packages.
Course requirements also include a programming assignment, which is
organized as a competition (see above).
The following are required to pass the course:
- 60% of exercise assignments solved. (70% -> +1 exam point; 80% -> +2
exam points; 90% -> +3 exam points).
- Project assignment, which is organized in the form of a pattern recognition
- Written exam (max. 30 pts; 15 to pass).
- Many topics of the course are also covered in Hastie et al., The
Elements of Statistical Learning, Springer 2009. Free PDF here.
- Scikit-learn documentation.
- If you forgot to register for the course, send email to the teacher (firstname.lastname@example.org)
- Unfortunately, no special arrangements are possible due to the size of the course.
- If you are unable to attend ANY exercises, request a separate assignment via email (email@example.com).