Washington University in St. Louis
Department of Computer Science and Engineering


CSE 417T: Introduction to Machine Learning

Spring 2019

ANNOUNCEMENTS

OVERVIEW

This course is an introduction to machine learning, focusing on supervised learning. We will cover the mathematical foundations of learning, a number of important techniques for classification and regression, including linear and logistic regression, neural networks, nearest neighbor techniques, kernel methods, decision trees, and ensemble methods, as well as some material on fairness, accountability, transparency, and ethics in machine learning. Note that the material in this course is a prerequisite for CSE 517A, the graduate level machine learning class. The overlap with CSE 511A (Artificial Intelligence) is minimal.

STAFF

Instructor: Sanmay Das
Office: Jolley 512
Office hours: Thursdays from 1-2, and by appointment.

TAs: There are several TAs for the class. Amanda Kube (amanda.kube at wustl), the graduate assistant to the instructor, will be the head TA and will conduct various recitation sessions as needed. The TAs will hold regular office hours (to be scheduled), grade homeworks, and answer questions on Piazza. The complete roster is as follows:
TA Email (at wustl.edu unless otherwise specified) Office Hours
Amanda Kube (Graduate Assistant to the Instructors) amanda.kube Mon 1-2 PM (Lopata 201)
Blake Bordelon blake.bordelon Thu 9-10 AM (Lopata 103)
Eric Cai ecai Tue 5-6 PM (Lopata 201)
Sam Griesemer samgriesemer at gmail dot com Tue 1-2 PM (Cupples I 216)
Jiaqi Hu hu.jiaqi Sun 3-4 PM (Lopata 103)
Feiran Jia feiran.jia Thu 1-2 PM (Cupples I 216)
Guancheng Jiang guancheng Wed 3-4 PM (Lopata 302)
Adam Kern adam.kern Wed 6-7 PM (Cupples I 216)
Jonathan Park jongwhan Thu 4-5 PM (Lopata 202)
Lexie Sun sunce Fri 2-3 PM (Lopata 103)
Cong Wang cwang41 Wed 11 AM-12 Noon (Lopata 202)
Sijia Wang sijiawang Mon 6-7 PM (Lopata 302)
Tong Wu tongwu Tue 3-4 PM (Lopata 103)
Shaohua Zhang shaohua Fri 9-10 AM (Lopata 103)
Louise Zhu bingluzhu Sat 1-2 PM (Sever 102)

The complete TA office hour schedule is as follows:
Sundays 3-4 PM (Lopata 103)
Mondays 1-2 PM (Lopata 201); 6-7 PM (Lopata 302)
Tuesdays 1-2 PM (Cupples I 216); 3-4 PM (Lopata 103); 5-6 PM (Lopata 201)
Wednesdays 11 AM-12 Noon (Lopata 202); 3-4 PM (Lopata 302); 6-7 PM (Cupples I 216)
Thursdays 9-10 AM (Lopata 103); 1-2 PM (Cupples I 216); 4-5 PM (Lopata 202)
Fridays 9-10 AM (Lopata 103); 2-3 PM (Lopata 103)
Saturdays 1-2 PM (Sever 102)

POLICIES

Detailed policies are in the official syllabus. A few points to highlight: please read and understand the collaboration policy and the late day policy. There will be two in-class exams, each covering approximately half the course material, and no separate final exam.

TEXTBOOKS

The main course textbook is: We also plan to cover some sections of the following book:

PREREQUISITES

CSE 247, ESE 326 (or Math 320), Math 233, and Math 309 (can be taken concurrently) or equivalents. If you do not have a solid background in calculus, probability, and computer science through a class in data structures and algorithms then you may have a hard time in this class. Matrix algebra will be used and is fundamental to modern machine learning, but it's OK to take that class concurrently.
--> --> --> -->
Date Topics Readings Assignments
Jan 15 Introduction. Course policies. Course overview. Slides; AML 1.1, 1.2.
Jan 17 The perceptron learning algorithm. Is learning feasible? AML Section 1.1.2, Problem 1.3, Section 1.3.1
Jan 22 Generalizing outside the training set. Error and noise. AML 1.3, 1.4
Jan 24 Infinite hypothesis spaces. VC dimension. AML 2.1.1-2.1.3 HW1 out
Gradescope instructions
Jan 29 The VC generalization bound (Prof. Ho). AML 2.1.4, 2.2
Jan 31 No in-class lecture. AI for Social Good by Prof. Milind Tambe at AAAI
Feb 5 The bias-variance tradeoff. AML 2.3.1
Feb 7 Bias-variance tradeoff, continued. Learning linear models with noisy data. AML 2.3.2, 3.1, 3.2
Feb 12 Logistic regression and gradient descent. AML 3.3 HW2 out
Feb 14 Nonlinear transformations. Overfitting. Regularization. AML 3.4, 4.1, 4.2
Feb 19 Regularization contd. Validation AML 4.2, 4.3 Malik Magdon-Ismail's slides on validation
Feb 21 Overfitting (reprise); Occam's razor, sample selection bias, and data snooping. AML 4.1, Chapter 5
Feb 26 Exam review.
Feb 28 In-class exam #1
Mar 5 Intro to decision trees. Tom Mitchell, Machine Learning Ch3; CASI 8.4
Mar 7 Decision trees, contd. Same as last lecture. HW3 out
Mar 19 Midterm discussion. Pruning. Bagging. CASI 17.1
Mar 21 Random forests. Boosting. CASI 17.1, 17.5, AdaBoost training error theorem proof
Mar 27 Lecture by Amanda Kube on causal inference. Slides are in Piazza resources
Mar 29 Final thoughts on boosting. Nearest neighbor methods AML eChapter 6.1-6.2 HW4 out
Apr 2 Efficient nearest neighbor search (k-d trees and LSH); Wikipedia articles on k-d trees and LSH. AML eChapter 6.3, 6.3.1.
Apr 4 RBF Networks. Fairness in ML. AML eChapter 6.3.1, 6.3.2. Propublica article on risk assessments,
The Measure and Mismeasure of Fairness: A Critical Review of Fair Machine Learning (Corbett-Davies and Goel)		HW5 out
Apr 9 Fairness in ML continued. Readings for HW5 on Piazza resources
Apr 11 Support vector machines. AML eChapter 8.1, 8.2, 8.4 HW6 out (due Apr 20)
Apr 16 Multilayer neural networks and backpropagation. AML eChapter 7.1, 7.2
Apr 18 Text analytics Slides