This is a repo for a 2021 course at the Department of Cosmic Research, MSU. The course covers the very basic concepts of ML, it is obligatory for 4th year students, and for MS students it is an elective course.
- The course consists of 12 lectures and 12 seminars and 5-6 homeworks.
- All lectures and seminars will be held online via Zoom and probably recorded (Zoom screen capture).
- PDFs with lecture slides will be published here: https://disk.yandex.ru/d/_tj5jTusnb0CEQ
- All homeworks are to be submitted online.
!!There will be no exams, final mark is based on homeworks only!!
- email
[email protected] - telegram news channel: https://t.me/joinchat/Or_wcTZUHo84ZTNi
- telegram chat room: https://t.me/joinchat/n3h9HxfzLDQ1MWNi
The timetable may be change.
- Lectures: 13:00 - 14:30 (UTC+3)
- Seminars: 14:40 - 16:00 (UTC+3)
Presentations and videos can be downloaded from here: https://disk.yandex.ru/d/_tj5jTusnb0CEQ
| # | Date | Lecture | Seminar | Homework |
|---|---|---|---|---|
| 1 | 2021-09-09 | Introduction to the topic | How to manage this course | HW#1 |
| 2 | 2021-09-16 | Metric Algorithms, kNN | Classification Quality Metrics | - |
| 3 | 2021-09-23 | Decision Trees | How to tune hyperparameters | - |
| 4 | 2021-09-30 | Linear Models | Invited speaker: Vasily Lavrov. Intro to Data Visualization | HW#2 |
| 5 | 2021-10-07 | SVM | PCA and SVD | - |
| 6 | 2021-10-14 | Regularized Linear Models | Useful web-pages for your DS career | HW#3 |
| 7 | 2021-10-21 | Time Series | Hands-on practice: design your alerting system | - |
| 8 | 2021-10-28 | Text Classification | Invited Speakers: how to apply and get hired | HW#4 |
| 9 | 2021-11-11 | Bayes approach, EM | TBA | - |
| 10 | 2021-11-18 | Clustering | Clustering hands-on practice: is your password weak? | HW#5 |
| 11 | 2021-11-25 | Ranking | TBA | - |
| 12 | 2021-12-02 | Gradient Boosting | TBA | HW#6 |
| 13 | 2021-12-09 | No Lecture Here | TBA | - |
Each homework has its deadline. Submission after deadline will reduce points.
| # | Name | Date Published | Deadline | Link |
|---|---|---|---|---|
| 1 | HW#1 Intro to numpy and pandas | 2021-09-10 | 2021-09-29 23:00:00 +03:00 | [link] |
| 2 | HW#2 Basic feature selection and hyperparameter tuning | 2021-09-30 | 2021-10-20 23:00:00 +03:00 | TBA |
| 3 | HW#3 Kaggle Comptetition: Classifiaction | 2021-10-14 | 2021-10-24 23:00:00 +03:00 / 2021-11-03 23:00:00 +03:00 | TBA |
| 4 | HW#4 TextClassification | 2021-10-28 | 2021-11-17 23:00:00 +03:00 | TBA |
| 5 | HW#5 Kaggle Competition: Regression | 2021-11-18 | 2021-11-28 23:00:00 +03:00 / 2021-12-01 23:00:00 +03:00 | TBA |
- Deadlines: HW#1 - 2 weeks, HW#2 and HW#4 - 3 weeks
- Submission after the deadline: each score is multiplied by 0.5
- Late submissions are not allowed in HW#3 and HW#5
- Submission rules for HW#3: 10 days to beat the baseline, 10 days to get your best score. No late submissions
- Submission rules for HW#5: 10 days to beat the baseline, 10 days to get your best score. No late submissions
- Everyone starts with 0.
- Beating Baseline before Medium Deadline +5
- 1st place +17
- 2nd place +15
- 3rd place +13
- Places 4-6 +10
- Places 7-9 +9
- Places 10-12 +8
- Places 13-15 +7
- Places 16-18 +6
- And so on
This course was inspired by
- The course by Konstantin Vorontsov from the Coursera [link]
- MIT OpenCourseWare Machine Learning course: Rohit Singh, Tommi Jaakkola, and Ali Mohammad. 6.867 Machine Learning. Fall 2006. Massachusetts Institute of Technology: MIT OpenCourseWare, https://ocw.mit.edu. License: Creative Commons BY-NC-SA.
- Data Mining In Action: [link]
- Author's personal experience in Data Science obtained from Yandex School of Data Science [link], working at Yandex [link], Lensa [link] and various ML-projects.