CPSC 340 and 532M - Machine Learning and Data Mining (Fall 2019)

Lectures Sections (beginning September 4):

2-3pm (Monday/Wednesday/Friday in Chemisty D300)
4-5pm (Monday/Wednesday/Friday in MacMillan 166)

Instructor: Mark Schmidt
Instructor office hours will start at the end of classes Wednesdays, and then move to the following location:

3-4pm (ICICS 146*)
5pm-6pm (ICICS X836**, will end early if no students remain)

*Will be held in ICICS 193 on Sep 4, Oct 16, and Nov 13.
**Will be held in ICICS 193 on Sep 11 and Sep 25, and cancelled on October 2.

Tutorials (beginning September 9):

1-2pm (Mondays in DMP 201)
5-6pm (Mondays in DMP 101)
330-430pm (Tuesdays in DMP 201)
430-530pm (Tuesdays in DMP 201)
12-1pm (Wednesdays in DMP 201)
9-10am (Thursdays in DMP 201)
10-11am (Thursdays in DMP 201)
11-12pm (Thursdays in DMP 201)
1-2pm (Thursdays in DMP 201)

Teaching assistants: Sarah Elhammadi, Dylan Green, Nam Hee Kim (Head TA), Frederik Kunstner, Ke (Mark) Ma, Lironne Kurzman, Benjamin Paul-Dubois-Taine, Michael Przystupa, Shahriar Shayesteh, Betty Shea, Karl Slakov, Yihan (Joey) Zhou
TA office hours (all in Demco Learning Centre):

Synopsis: We introduce basic principles and techniques in the fields of data mining and machine learning. These are some of the key tools behind the emerging field of data science and the popularity of the `big data' buzzword. These techniques are now running behind the scenes to discover patterns and make predictions in various applications in our daily lives. We'll focus on many of the core data mining and machine learning technlogies, with motivating applications from a variety of disciplines.

Registration: Undergraduate and graduate students from any department are welcome to take the course. Undergraduate students should enroll in CPSC 340 while graduate students should enroll in CPSC 532M (which has an extra small project component). Below are more details on registration for each course:

The majority of the seats in 340 are reserved for UBC computer science majors. For other students, to enroll in the course you need to sign up for the wait list. Note that last year all students on the wait list were ultimately accepted into the course (but we did not have room for auditors.)
Two-thirds of the seats in 532M will initially be reserved for UBC computer science grad students, while the remaining seats are open to graduate students from any department. Note that you will likely get a warning about prerequisities when you register (which you can ignore), but depending on your department you may later be contacted and asked to show how you satisfy the relevant prerequisites. As with 340, if 532M becomes full then signing up for the waiting list is the only way to enroll in the course. Last year all students on the wait list were ultimately accepted into the course.

Starting in the second week of classes, we'll have weekly tutorials run by the TAs. These will do things like go through provided assignment code, review background material, review big concepts, and/or do exercises. You can register for particular tutorial sections if you want to save a seat at a particular time, but note that you do not need to register in a tutorial section.

Prerequisites:

Basic algorithms and data structures (CPSC 221, or both of CPSC 260 and EECE 320 as well as one of CPSC 210, EECE 201, or EECE 309).
Linear algebra (one of MATH 152, 221, or 223).
***Probability (one of STAT 241, STAT 251, ECON 325, ECON 327, MATH 302, STAT 302, or MATH 318).
Multivariate calculus (one of MATH 200, 217, 226, 253, or 263).

***Students who completed STAT 200, STAT 203, PSYC 218, PSYC 278, BIOL 300 or COMM 291 with a grade of at least 72%, prior to 2019W1 will be allowed to take CPSC 340 in 2019W, provided they have all of the other stated prerequisites. Students who do not meet these requirements should consider taking CPSC 330, a new course on applied machine learning.

Textbook: There is no required textbook for the class. A introductory book that covers many (but not all) the topics we will discuss is the Artificial Intelligence book of Rusell and Norvig (AI:AMA) or the Artificial Intelligence book of Poole and Mackworth (you may need these for other classes). More advanced books include The Elements of Statistical Learning (ESL) by Hastie et al., Murphy's Machine Learning: A Probabilistic Perspective (ML:APP) which can be accessed through the library here, and Bishop's Pattern Recognition and Machine Learning (PRML). For books with a bigger focus on data mining, see Introduction to Data Mining (IDM) and Mining of Massive DataSets.

Related Courses: The most related course is CPSC 330: Applied Machine Learning. This course has fewer prerequisities and covers some of the same material, but focuses more on applications rather than understanding ML ideas in depth.

Related courses in statistics include: STAT 305, STAT 306, STAT 406, STAT 460, STAT 461 (as well as EOSC 510). A discussion of the difference between CPSC 340 and these various STAT classes written by a former student (Geoff Roeder) is available here.

Grading:

340: Assignments 30%, Midterm 20%, Final 50%.
532M: Assignments 25%, Midterm 15%, Final 40%, Project 20%.

Piazza for course-related questions.

List of topics

We will roughly cover the following topics:

Data representation and summarization.
Supervised learning with frequencies and distances.
Data clustering, outlier detection, and association rules.
Linear prediction, regularization, and kernels.
Latent-factor models and collaborative filtering.
Neural networks and deep learning.

Lectures, Assignments, Related Readings, and Links

Date	Slides	Related Readings and Links	Homework and Notes
Wed Sep 4	Motivation and Syllabus	What is Machine Learning? Machine Learning Rise of the Machines Talking Machine Episode 1
Fri Sep 6	Exploratory Data Analysis	Gotta Catch'em all Why Not to Trust Statistics Visualization Types Google Chart Gallery Other Tools	Assignment 1 a1.zip a1.tex
Mon Sep 9	Decision Trees	A Visual Introduction to Machine Learning, Decision Trees Entropy AI:AMA 18.2-3, ESL: 9.2, ML:APP 16.2	Big-O Notes Julia Commands
Wed Sep 11	Fundamentals of Learning	7 Steps of Machine Learning IID Cross-validation Bias-variance No Free Lunch AI: AMA 18.4-5, ESL 7.1-7.4, 7.10, ML:APP 1.4, 6.5	Course Notation Guide
Fri Sep 13	Probabilistic Classifiers	Conditional probability (demo) Naive Bayes Probabilities and Battleship ESL 4.3, ML: APP 2.2, 3.5, 4.1-4.2	Assignment 1 due Probability Notes Probability Slides
Mon Sep 16	Non-Parametric Models	K-nearest neighbours Decision Theory for Darts Norms AI: AMA 18.8, ESL 13.3, ML:APP 1.4	Assignment 2 a2.zip a2.tex
Wed Sep 18	Ensemble Methods	Ensemble Methods Random Forests Empirical Study Kinect AI: AMA 18.10, ESL: 7.11, 8.2, 15, 16.3, ML: APP 6.2.1, 16.2.5, 16.6
Fri Sep 20	Clustering	Clustering K-means clustering (demo) K-Means++ (demo) IDM 8.1-8.2, ESL: 14.3
Mon Sep 23	More Clustering	DBSCAN (video, demo) Hierarchical Clustering Phylogenetic Trees IDM 8.4
Wed Sep 25	Outlier Detection	Empirical Study IDM 8.3, ESL 14.3.12, ML:APP 25.5
Fri Sep 27	Finding Similar Items (Bonus Lecture)	MMD Chapter 3	Assignment 2 due
Mon Sep 30	Least Squares	Linear Regression (demo, 2D data, 2D video) Least Squares Essence of Calculus Partial Derivative Gradient ESL 3.1-2, ML:APP 7.1-3, AI:AMA 18.6	Assignment 3 a3.zip a3.tex Calculus Notes
Wed Oct 2	Nonlinear Regression	Why should one learn machine learning from scratch? Essence of Linear Algebra Matrix Differentiation Fluid Simulation (video) ESL 5.1, 6.3	Linear Algebra Notes Linear/Quadratic Gradients
Fri Oct 4	Gradient Descent	Gradient Descent Convex Functions
Mon Oct 7	Robust Regression	ML:APP 7.4
Wed Oct 9	Feature Selection	Genome-Wide Association Studies AIC, BIC ESL 3.3 , 7.5-7
Fri Oct 11	Regularization	ESL 3.4., ML:APP 7.5, AI:AMA 18.4	Assignment 3 due
Wed Oct 16	More Regularization	RBF video RBF and Regularization video ESL 6.7, ML:APP 13.3-4
Thu Oct 17	Midterm
Fri Oct 18	Linear Classifiers	Perceptron ESL 4.5, ML:APP 8.5
Mon Oct 21	More Linear Classifiers	Support Vector Machines ESL 4.4, 12.1-2, ML:APP 8.1-3, 9.5 14.5, AI:AMA 18.9	Assignment 4 a4.zip a4.tex
Wed Oct 23	Feature Engineering	Gmail Priority Inbox
Fri Oct 25	Convolutions
Mon Oct 28	Kernel Trick	ESL 12.3, ML:APP 14.1-4
Wed Oct 30	Stochastic Gradient	Stochastic Gradient ML:APP 8.5
Fri Nov 1	Boosting	AdaBoost (video) XGBoost (video) ML:APP 16.4	Assignment 4 due Max and Argmax Notes
Mon Nov 4	MLE and MAP	Maximum Likelihood Estimation ML:APP 9.3-4	Assignment 5 a5.zip a5.tex
Wed Nov 6	Principal Component Analysis	Principal Component Analysis ESL 14.5, IDM B.1, ML:APP 12.2
Fri Nov 8	More PCA	Making Sense of PCA SVD Eigenfaces
Wed Nov 13	Sparse Matrix Factorization	Non-Negative Matrix Factorization (original - access from UBC) ESL 14.6, ML: APP 13.8
Fri Nov 15	Recommender Systems	Recommender Systems Netflix Prize	Assignment 5 due
Mon Nov 18	Multi-Dimensional Scaling	Nonlinear Dimensionality Reduction t-SNE demo ESL 14.8-9, IDM B.2	Assignment 6 a6.zip a6.tex
Wed Nov 20	Deep Learning	Google Video What is a Neural Network? Interactive Guide ML:APP 16.5, ESL 11.1-4, AI: AMA 18.7
Fri Nov 22	More Deep Learning	Fortune Article Deep Learning References Alchemy ML:APP 28.3, ESL 11.5
Mon Nov 25	Convolutional Neural Networks	Convolutional Neural Networks ML:APP 28.4, ESL 11.7
Wed Nov 27	2pm: Semi-Supervised Learning (CHEM D300) 4pm: More CNNs (MCML 166)	Semi-Supervised Learning Label Propagation at Google AlexNet
Fri Nov 29	2pm: PageRank (CHEM D300) 4pm: Number of Iterations (MCML 166)	PageRank Slides PageRank Math/Code, ESL 14.10, AI: AMA 22.3 Non-convex PL Inequality	Assignment 6 due

Mike's Demos

In semesters where Mike Gelbart teaches the course, he uses a variety of Python notebooks. Julia versions of these notebooks are available here.

Related courses that have online notes

Machine Learning and Data Mining (UBC 2012)
Introduction to Machine Learning (Alberta)
Practical Machine Learning (Berkeley)
Machine Learning (MIT)
Machine Learning (CMU)
Course in Machine Learning (Maryland)
Principals of Knowledge Discovery in Data (Alberta)
Data Mining (CMU)

Mark Schmidt > Courses > CPSC 540