COMS20011_2020

Data-driven Computer Science UoB

Staff

• Laurence Aitchison [laurence.aitchison@bristol.ac.uk] (unit director)
• Majid Mirmehdi [m.mirmehdi@bristol.ac.uk]

Teaching Assistants

Holly Milllea | Amirhossein Dadashzadeh | Faegheh Sardari | Jonathan Munro | Vangelis Kazakos | Zhaozhen Xu

Structure

Lecture videos for a week will be released on Monday and posted here. Please take a look at them promptly!

The coursework will be 40% of the mark and the exam will be 60% (deadlines TBA).

TA led sessions on Thursday 3-5 are the main route for feedback on all aspects of the course: lectures, labs and coursework. These will start on 11th Feb, and you will be assigned a group by email before then. They will be hosted through the public Teams group [grp-COMS20011_2020] (if the link doesn't work, just search for "grp-COMS20011_2020").

There are lecturer-led Q&A sessions on Mondays at 4pm. The first of these (Feb 1st) will be TA led, to help getting IT set up for labs.

There will be "lab" exercises released in the "lab" folder. Please do them promptly and bring any questions to the TA-led sessions: the coursework is heavily based on the labs!

Mathematical background material

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.

Probability and statistics

• A Modern Introduction to Probability and Statistics, Understanding Why and How (Dekking et al.)
• [MIT OpenCourseWare]
• [Khan Academy] (from Probability)

Calculus:

• [MIT OpenCourseWare]
• [Khan Academy]

Linear Algebra:

• Linear Algebra for Everyone (Gilbert Strang)
• [Khan Academy]

All of the above

• [Mathematics for Machine Learning]

Weekly lecture material

Week 13: 01/02/2021 (Majid)

Lecture	Duration	video	slides
MM01. Intro to COMS20111 - very fishy	14:44	[Stream link]	[pdf]
MM02. Intro - Part 2 - example projects	12:19	[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 Updated - New Q 12/02/21	-	Self/Group study	[pdf]
Problem Sheet Updated - New Q/A 12/02/21	-	Answers	[pdf]
Q&A Session	60:00	[Stream link]	-

(Week 14): 08/02/2021 (Laurence)

Lecture	video	slides
1. Maximum likelihood for a coin	[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	[Stream link]	[notebook 2]
Problem Sheet W14		[pdf]
Problem Sheet W14 Solution Explanation		[pdf]
Q&A Session	[Stream link]	-

(Week 15): 14/02/2021 (Laurence)

Lecture	video	slides
1. Linear regression examples	[Stream link]	[notebook 2]
2. Overfitting	[Stream link]	[notebook 3]
3. Cross-validation	[Stream link]	[notebook 3]
4. Regularisation	[Stream link]	[notebook 3]
Problem Sheet W15		[notebook]
Problem Sheet W15 Solution Explanation		[pdf]
Q&A Session	[Stream link]	-

(Week 16): 14/02/2021 (Laurence)

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]
4. Bayesian classification	[Stream link]	[notebook 4]
Problem Sheet W16		[notebook]
Q&A Session	[Stream link]	-

(Week 18): 8/03/2021 (Laurence)

Lecture	video	slides
1. Clustering vs classification	[Stream link]	[notebook 5]
2. K-means clustering	[Stream link]	[notebook 5]
3. EM for Gaussian mixture models	[Stream link]	[notebook 5]
4. Objective for EM [Non-examinable]	[Stream link]	[notebook 5]
Problem Sheet W18		[notebook]
Q&A Session (lots about CW!)	[Stream link]	-

Notebook pdfs

I have printed the Notebooks as pdfs. Note that this really doesn't work well, as many of the interactive plots can't be printed.

Notebook
[notebook 1]
[notebook 2]
[notebook 3]
[notebook 4]

Week 19: 15/03/2021 (Majid)

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]]
Q&A Session	-	[[Stream link]]	-

Week 20: 22/03/2021 (Majid)

Lecture	Duration	video	slides
MM09. 2D Fourier Transform	-	[[Stream link]]	[[pdf]]
MM10. Frequency Features	-	[[Stream link]]	[[pdf]]
Problem Sheet MM03	-	Self/Group study	[[pdf]]
Problem Sheet MM03	-	Answers	[[pdf]]
Q&A Session	-	[[Stream link]]	-

Week 20: 29/03/2021 (Majid)

Lecture	Duration	video	slides
MM11. More on Features	-	[[Stream link]]	[[pdf]]
MM12. Convolutions	-	[[Stream link]]	[[pdf]]
Problem Sheet MM04	-	Self/Group study	[[pdf]]
Problem Sheet MM04	-	Answers	[[pdf]]
Q&A Session	-	[[Stream link]]	-