Data-driven Computer Science UoB
- Laurence Aitchison [[email protected]] (unit director)
- Majid Mirmehdi [[email protected]]
- Zhaozhen Xu, Sheng Wang, Frederick Turner, Amirhossein Dadashzadeh, Faegheh Sardari, Xinyu Yang.
Lecture videos for a week will be released on Monday and posted here. Please take a look at them promptly!
The Teams Team is "COMS20011: Data-Driven Computer Science 2021/22 (TB-2, A)". You should have access already through the "Teams" panel. If not, please get in touch with the school office!
The course is assessed 100% by exam.
The routes for feedback are:
- Lecturer led QA sessions on Tuesdays at 2pm
- TA led sessions on Thursday 1-2 (group 1) or 2-3 (group 2) (you should have one of these slots in your calendar). You can ask about all aspects of the course lectures, labs and exams. We have a room for these (MVB 2.11), and some support will be available in the room, and some online, through Teams (which can also be accessed in the room). The first lab session is on 27th Feb.
There will be "lab" exercises released in the "lab" folder.
As the course has only run in its present form for one year, we only have one past exam paper with answers
Important: these are not pre-requisites! Please don't try to look at all of the material! They're intended as supplements to the first-year maths courses to help clear up specific issues with the derivations in the course. Feel free to raise an issue/pull-request if you have recommendations for other resources.
- Probabilistic Machine Learning: An Introduction pdf (Parts I - III are most relevant. Obviously that's quite a bit, ranging from fundamentals to deep learning. But different people will find different bits most relevant).
- A Modern Introduction to Probability and Statistics, Understanding Why and How (Dekking et al.)
- [MIT OpenCourseWare]
- [Khan Academy] (from Probability)
- Linear Algebra for Everyone (Gilbert Strang)
- [Khan Academy]
Lecture | Duration | video | slides |
---|---|---|---|
MM01. Intro to COMS20111 - very fishy | 14:35 | [Stream link] | [pdf] |
MM02. Intro - Part 2 - example projects | 10:48 | [Stream link] | [pdf] |
MM03. Data Acquisition - Sampling - Acquisition | 10:38 | [Stream link] | [pdf] |
MM04. Data Characteristics - Distance Measures | 15:55 | [Stream link] | [pdf] |
MM05. Data Characteristics - Covariance - Eigen Analysis - Outliers | 20:50 | [Stream link] | [pdf] |
Problem Sheet | - | Self/Group study | [pdf] |
Problem Sheet | - | Answers | [pdf] |
Q&A Session Tuesday Feb 1st | 60:00 | [Stream link] | - |
Lecture | video | slides |
---|---|---|
1. Maximum likelihood for a coin (ignore references to cw; 100% exam this year) | [Stream link] | [notebook 1] |
2. Bayes for a coin | [Stream link] | [notebook 1] |
3. Intro to supervised learning | [Stream link] | [notebook 2] |
4. Linear regression derivation (non-examinable) | [Stream link] | [notebook 2] |
Problem Sheet | [pdf] | |
Problem Sheet Explanation | [pdf] | |
Q&A Session | [Stream link] |
Lecture | video | slides |
---|---|---|
1. Regression: examples | [Stream link] | [notebook 2] |
2. Regression: Overfitting | [Stream link] | [notebook 3] |
3. Regression: Cross-validation | [Stream link] | [notebook 3] |
4. Regression: Regularisation | [Stream link] | [notebook 3] |
Problem Sheet | [notebook] | |
Problem Sheet Explanation | [pdf] | |
Q&A Session (no sound for the first two mins) | [Stream link] |
Lecture | video | slides |
---|---|---|
1. Logits parameterisation | [Stream link] | [notebook 4] |
2. Gradient descent + overfitting | [Stream link] | [notebook 4] |
3. KNN/WNN and nearest centroids | [Stream link] | [notebook 4] |
Problem Sheet W16 | [notebook] | |
Problem Sheet Explanation | [pdf] | |
Q&A Session | [Stream link] |
Lecture | video | slides |
---|---|---|
1. Bayesian classification | [Stream link] | [notebook 4] |
2. Clustering vs classification | [Stream link] | [notebook 5] |
3. K-means clustering | [Stream link] | [notebook 5] |
Problem Sheet W17 | [notebook] | |
Problem Sheet Explanation | [pdf] | |
Q&A Session | [Stream link] |
Lecture | video | slides |
---|---|---|
1. EM for Gaussian mixture models | [Stream link] | [notebook 5] |
2. Objective for EM [Non-examinable] | [Stream link] | [notebook 5] |
Problem Sheet W18 | (No problem sheet this week) | |
Q&A Session | [Stream link] |
Lecture | Duration | video | slides |
---|---|---|---|
MM06. Signals & Frequencies | 13:26 | [Stream link] | [pdf] |
MM07. Fourier Series | 10:28 | [Stream link] | [pdf] |
MM08. 1D Fourier Transform | 17:18 | [Stream link] | [pdf] |
Problem Sheet MM02 | - | Self/Group study | [pdf] |
Problem Sheet MM02 | - | Answers | [pdf] |
Code to play with | - | sines.py | |
Q&A Session | 45:40 | [Stream link] | - |
Lecture | Duration | video | slides |
---|---|---|---|
MM09. 2D Fourier Transform | 14:45 | [Stream link] | [pdf] |
MM10. Frequency Features | 19:16 | [Stream link] | [pdf] |
Problem Sheet MM03 | - | Self/Group study | [pdf] |
Problem Sheet MM03 | - | Answers | [pdf] |
Q&A Session | 23:58 | [Stream link] | - |
Lecture | Duration | video | slides |
---|---|---|---|
MM11. More on Features | 20:53 | [Stream link] | [pdf] |
MM12. Convolutions | 20:02 | [Stream link] | [pdf] |
Optional Playthings | *** | [sobel.py] | [FFT.py] |
Problem Sheet MM04 | - | Self/Group study | [pdf] |
Problem Sheet MM04 | - | Answers | [pdf] |
Q&A Session | 36:40 | [Stream link] | - |
- [1pm on May 5th: ONLINE (this is during the lab session) ] - Fun Quiz online on Teams - the link to join the Quiz is https://kahoot.it/ and will be posted in the usual channel (grp-COMS20011_2021). Join our meeting in the channel first at 1pm on May 5th. The Quiz will includes multiple choice questions on Majid's part of the unit (but rather simple as time will be limited) as well as some other fun questions!