| Top | Policies | Calendar | Materials | Contributors |
Machine learning is the study of algorithms that improve automatically through experience. This is an overview class that covers the fundamentals: classification, regression, and model training and evaluation across a variety of approaches, both old and new.
For this course, it is extremely helpful to be familiar with probability, statistics, python programming and matrix math. We expect students to come prepared with those skills. Here are some prerequisite materials to get up to speed, in case you need a refresher.
Section 1: Mon Wed 9:30AM - 10:50AM Frances Searle Building 2107
Section 2: Tue Thu 8:00AM - 9:20AM Harris Hall 107
Videos lectures will be available on Canvas.
These hours are subject to change depending on demand. A calendar view of these office hours is available here.
Bryan Pardo: Tue Thu 9:40am - 11:00am via Zoom Go here to book a 10 minute meeting.
Zach Wood-Doughty: Mon Fri 1:30-3pm via Zoom
Hugo Flores Garcia: Tue Thu @ 5-6pm via Zoom
Patrick O’Reilly: Sat 2-4pm via Zoom
Alex Tai: Sun 9-11am in Tech EG20
Boaz Cogan: Mon Wed 8-9am in Tech EG20
Noah Schaffer: Fri 9-11am in Tech EG20
Aldo Aguilar: Mon Wed 5-6pm via Zoom
Garphy Tam: Tue Wed 3-4pm via Zoom
Simran Gadkari: Tue Fri 12-1pm via Zoom
Utkarsh Mishra: Wed Thu 6-7pm via Zoom
Grading: There are 6 assignments, each worth 20 points. We will calculate your grade as follows: Throw out the lowest grade, average the other 5 and scale to the range 0-100. You’re graded on a basis of 100 points. 93-100 is an A, 90 - 92 is an A-, 87-89 is a B+, 83-86 is a B, 80-82 is a B-…and so on. There is no curve.
Extra Credit: The first assignment is extra-credit, extremely easy and worth 20% of the total points. No special additional extra credit will be given. Remember, we’re also dropping your lowest-graded assignment. Practically speaking, this means you can entirely skip any one assignment and still get 100 out of 100. This can be helpful if you are ill at some point or have other commitments.
When and Where to Submit Assignments: Free responses are due on Canvas by 11:59pm on the due date. Code is due in your individual GitHub Classroom code repository, also at 11:59pm on the due date.
Late Policy: If there is nothing on Canvas by 11:59pm on the due date, the free response grade is 0. If there is nothing in your github repository by 11:59pm on the due date, the coding grade is 0. The most recent code on github at 11:59pm on the due date is the code we will grade. The most recent submission in Canvas at that point, is the one we grade. A good approach is to continually check in and push to GitHub as you work. Also, put up a “safety” submission on Canvas with what you currently have, an hour prior to the deadline.
Cheating & Academic Dishonesty: Do your own work. This includes free response answers and code. Penalties include failing the class and can be more severe than that. If you have a question about whether something may be considered cheating, ask, prior to submitting your work. We will be checking for code duplication. Academic dishonesty will be dealt with as laid out in the student handbook. We will be using MOSS for checking on academic dishonesty.
Attendance is not graded. You are not required to attend. Videos of the Tuesday/Thursday lectures will be made available on Canvas.
COVID 19 and Health: Videos will be available for all course lecture content. If you have tested postive for COVID19 do not come to class. If you feel ill, do not come to class. Watch the video of the lecture instead. If anyone is seen in class without a mask covering their mouth and nose, class will be halted until that person either leaves or puts on a mask.
Announcements and discussions will take place on CampusWire. You can sign up for the page at that link using the sign-up code 5975.
This is a prediction of what will be covered in each week but the schedule is subject to change as the course progresses.
| Week | Due | Topic (click on topic for slides) | Lecturer |
|---|---|---|---|
| 09/20 | No class | ||
| Intro to ML | Pardo | ||
| 09/27 | HW 0 | P-norms and Distance Measures | Pardo |
| Nearest Neighbor Classifiers | Pardo | ||
| 10/04 | Linear and Polynomial Regression | Wood-Doughty | |
| Linear Discriminants | Wood-Doughty | ||
| 10/11 | HW 1 | Decision trees | Pardo |
| Measuring Error | Pardo | ||
| 10/18 | Gradient Descent | Pardo | |
| Multilayer Perceptrons and Backprop | Pardo | ||
| 10/25 | HW 2 | Convolutional Networks | Pardo |
| Convolutional networks continued | Pardo | ||
| 11/01 | Clustering | Wood-Doughty | |
| Expectation Maximization | Wood-Doughty | ||
| 11/08 | HW 3 | Graphical Models | Wood-Doughty |
| Graphical Models (continued) | Wood-Doughty | ||
| 11/15 | Hidden Markov Models | Wood-Doughty | |
| Causal Graphical Models | Wood-Doughty | ||
| 11/22 | HW 4 | Reinforcement Learning | Pardo |
| NO CLASS: THANKSGIVING | |||
| 11/29 | Fairness, accountability, transparency, and ethics | Wood-Doughty | |
| 12/01 | Current research topics in ML | O’Reilly | |
| 12/02 | Current research topics in ML | Garcia | |
| 12/08 | HW 5 | NO CLASS: Finals week |
The lecture slides will be made available on the course calendar above, as the course progresses.
Videos of the Tuesday/Thursday lectures will be made available on Canvas (via Panopto) on the same day they are recorded.
People who contributed to the design of this course and the homeworks:
Bryan Pardo, Prem Seetharaman, Bongjun Kim, Max Morrison, Ethan Manilow, Fatemeh Pishdadian, Lukas Justen, Oliver Coissart, Florian Schiffers, Sushobhan Ghosh, Ze Zhu, Zach Wood-Doughty, Hugo Flores Garcia, Patrick O’Reilly
| Top | Policies | Calendar | Materials | Contributors |