janetxinli / cpsc330

CPSC 330: Applied Machine Learning

UBC CPSC 330: Applied Machine Learning (2020W1)

This is the course homepage for CPSC 330: Applied Machine Learning at the University of British Columbia. You are looking at the current version (Sep-Dec 2020). An earlier version from Jan-Apr 2020 can be found here.

Instructor: Mike Gelbart

Important links

• syllabus / administrative info
• other course documents
• past exams
• Piazza (this is where all announcements will be made)
• class + office hours calendar
• Zoom links on Canvas

Lecture schedule

Live lectures: The lectures will be on Zoom. They can be joined through Canvas here. If you would like to join the lectures but cannot login to Canvas (presumably because you're not enrolled in the course) please email Mike and I will give you the link.

Lecture recordings: The lecture recordings can be accessed through the same Zoom page on Canvas here. From this page, navigate to the "Cloud Recordings" tab and you should see them there. The same lecture recordings will be posted here embedded in the schedule below.

#	Date	Topic	Recording	Related readings and links	vs. CPSC 340
	Sep 8	UBC Imagine Day - no class
1	Sep 10	Course intro	recording		n/a
		Dataset of the week: which CPSC 330 students like cilantro?
2	Sep 15	Decision trees	recording pw !?3niNc^		less depth
3	Sep 17	The fundamental tradeoff of ML	recording pw 90p@qbt4	About Train, Validation and Test sets	similar
		Dataset of the week: sentiment analysis of movie reviews
4	Sep 22	Logistic regression, word counts, predict_proba	recording pw ZsJZ#e29	Meaningless comparisons lead to false optimism in medical machine learning	less depth
5	Sep 24	Pipelines & hyperparameter optimization	recording pw pT2QVE*# and supplemental screencast		more depth
		Dataset of the week: Predicting income from census data
6	Sep 29	Overfitting the validation set & encoding categorical variables	recording pw 3J10UiO.		more depth
7	Oct 1	Imputation, scaling numeric features, ColumnTransformer	recording pw b.+DFR47		more depth
		Dataset of the week: detecting credit card fraud
8	Oct 6	Evaluation metrics for classification	recording	Damage Caused by Classification Accuracy and Other Discontinuous Improper Accuracy Scoring Rules, Classification vs. Prediction	more depth
9	Oct 8	Ensembles	recording		similar
			Dataset of the week: predicting housing prices
10	Oct 13	Linear regression, regression metrics	recording part 1, recording part 2		more depth on metrics, less on linear regression
11	Oct 15	Prediction intervals, feature importances	recording		new
12	Oct 20	Feature selection, midterm review	screencast	Feature selection article	less depth
	Oct 22	MIDTERM	study materials
13	Oct 27	Neural networks & computer vision	recording	But what is a Neural Network?	less depth
14	Oct 29	Distances and neighbours			less depth

Below this point is the schedule for future lectures (not yet updated for 2020W1)

#	Date	Topic	Recording	Related readings and links	vs. CPSC 340
15	Nov 3	Feature engineering, embeddings			new
16	Nov 5	Time series data		Humour: The Problem with Time & Timezones	new
17	Nov 10	Survival analysis		Calling Bullshit video 4.1, Medium article (contains some math)	new
18	Nov 12	Clustering			less depth
19	Nov 17	Outliers			different angle
20	Nov 19	Communicating your results		Communication in Data Science blog post; Calling BS videos Chapter 1 (5 video total)	new
21	Nov 24	Communicating your results, continued		Calling BS videos Chapter 6 (6 short videos, 47 min total)	new
22	Nov 26	Ethics		Calling BS videos Chapter 5 (6 short videos, 50 min total)	new
23	Dec 1	Model deployment			new
24	Dec 3	Leftovers; Conclusion

Homework schedule

#	Due Date	Associated lectures
1	Tue Sep 15 11:59pm	prerequisites
2	Mon Sep 21 11:59pm	2, 3
3	Mon Sep 28 11:59pm	4, 5
4	Mon Oct 12 11:59pm	6-9
5	Mon Oct 19 11:59pm	9-11
6	Mon Nov 9 11:59pm	12-15
7	Mon Nov 16 11:59pm	16-18
8	Mon Nov 23 11:59pm	whole course
9	Mon Nov 30 11:59pm	20-22

Attribution

Thank you to Tomas Beuzen and Varada Kolhatkar for significant contributions to the course materials.

License

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.