Introduction to Machine Learning

CMSC 478

Spring 2024

Contact Information

Instructor:

Tim Oates
Office: ITE-336
Email: oates@umbc.edu
Office hours: Monday 12:30 - 1:30

TA:

TBD
Office: TBD
Email: TBD
Office hours: TBD

We will rely on Slack for asynchronous communication. Please use this link to sign up for the class Slack. You can use Slack for discussions between yourselves and to ask questions of me or the TA. If you want to ensure that I see your post, please use @oates in it so that I get a notification. Most discussions tend to take place in the general channel, but feel free to ask me to create other channels.

Grading

Grades will be based on a midterm exam, a final exam, and seven homework assignments. The homeworks are crucial for solidifying what you learn in class.

The weights on the various items are as follows:

exams - 15% each, 30% total
homeworks - 10% each, 70% total

Grades will be assigned as follows based on your class average:

[90, 100] = A
[80, 90) = B
[70, 80) = C
[60, 70) = D
[0, 60) = F

Note that, for example, [80, 90) means the interval that includes 80 but not 90.

All assignments will be submitted to the TA via slack.

Late Policy

All assignments must be turned in by 11:59PM Eastern time on the date that they are due. I understand that students have many demands on their time that vary in intensity over the course of the semester. Therefore, you will be allowed 5 late days without penalty for the entire semester. You can turn in 5 different assignments one day late each, or one assignment 5 days late, and so on. Late days cannot be used for exams.

Once the late days are used, a penalty of 33% will be imposed for each day (or fraction thereof) an assignment is late (33% for one day, 66% for two, 100% for three or more). An assignment is late by one day if it is not turned in by 11:59PM Eastern on the day that it is due. It is late by two days if I do not have it by 11:59PM Eastern the following day, and so on. It is your responsibility to keep track of how many late days you have used.

Academic Honesty

By enrolling in this course, each student assumes the responsibilities of an active participant in UMBC’s scholarly community in which everyone’s academic work and behavior are held to the highest standards of honesty. Cheating, fabrication, plagiarism, and helping others to commit these acts are all forms of academic dishonesty, and they are wrong. Academic misconduct could result in disciplinary action that may include, but is not limited to, suspension or dismissal. To read the full Student Academic Conduct Policy, consult UMBC policies, or the Faculty Handbook (Section 14.3). For graduate courses, see the Graduate School website.

I will actively monitor all student work, as will the TA, for instances of academic misconduct. The penalty for any such misconduct on the first instance can be a zero on the assignment, depending on the severity. The penalty for the second instance can be an F in the class, depending on the severity.

Textbook

We will use a variety of online sources during this course.

The first few chapters of an interactive book I'm writing (see below)
ESL: Elements of Statistical Learning: https://web.stanford.edu/~hastie/ElemStatLearn/printings/ESLII_print12.pdf
ITILA: Information Theory, Inference and Learning Algorithms: http://www.inference.phy.cam.ac.uk/itprnn/book.pdf
UML: Understanding Machine Learning: From Theory to Algorithms: http://www.cs.huji.ac.il/~shais/UnderstandingMachineLearning/
CIML: A Course in Machine Learning: http://ciml.info/

Tools

In most cases you will be required to use python for homeworks because python has become the default language for machine learning at scale. Therefore, all of the examples that I do in class where we run an actual algorithm will be done using scikit-learn. A very easy way to get everything you may need is to install anaconda. It has python, scikit, and Jupyter notebooks for working with data and presenting results.

Syllabus

This syllabus is subject to small changes, but due dates and exam dates will not change. Note that for each topic there will be two sets of readings - some that everyone should do (marked all) and some that are optional (marked opt). You will only be held responsible for what is in the readings that everyone does, but if you want more information the optional readings are a good source.

Class	Date	Topic	Events/Readings
1	Mon 29 Jan	Course overview. What is machine learning?	Read Chapter 1 of my book
2	Wed 31 Jan	Perceptrons	Read Chapter 2 of my book
3	Mon 05 Feb	More percentrons	Homework 1 out. Here it is in HTML format if you do not already know how to open Python notebooks.
4	Wed 07 Feb	Loss functions, gradient descent	Read Chapter 3 of my book
5	Mon 12 Feb	More gradient descent	Homework 1 due; Homework 2 out - notebook and HTML
6	Wed 14 Feb	Logistic regression	Read Chapter 4 of my book
7	Mon 19 Feb	Clustering	Slides
8	Wed 21 Feb	Nearest neighbors	CIML chapter 3 through 3.4
9	Mon 26 Feb	Methodology	Slides Homework 2 due; Homework 3 out
10	Wed 28 Feb	More methodology
11	Mon 04 Mar	Reinforcement learning	Chapter 3 and 6.5 of the RL Book
12	Wed 06 Mar	More reinforcement learning	Slides - 01,03,04,05,06
13	Mon 11 Mar	A little more reinforcement learning	Homework 3 due
14	Wed 13 Mar	Kernel methods	Slides
	Mon 18 Mar	Spring Break
	Wed 20 Mar	Spring Break
15	Mon 25 Mar	Support vector machines
16	Wed 27 Mar	Midterm review
17	Mon 01 Apr	Midterm Exam on content of classes 1 - 14	Homework 4 out
18	Wed 03 Apr	Support vector machines
19	Mon 8 Apr	Neural networks (Slides)
20	Wed 10 Apr	More NNs	Homework 4 due Homework 5 out
21	Mon 15 Apr	Convolutional neural networks	Slides
22	Wed 17 Apr	More CNNs
23	Mon 22 Apr	Recurrent neural networks	Slides Homework 5 due Homework 6 out
24	Wed 24 Apr	More RNNs
25	Mon 29 Apr	Graphical models
26	Wed 01 May	Bayes nets (Slides)	Homework 6 due; Homework 7 out
27	Mon 06 May	More Bayes nets
28	Wed 8 May	Dimensionality reduction
29	Mon 13 May	Final exam review	Homework 7 due
	Mon 20 May	Final Exam 10:30AM - 12:30PM