Course Content
This page includes all the content for the course thus far. We will update this page with all lecture materials, readings, and homework as the class goes on.
Schedule and Main Content
This class has six main modules, two for each “pillar” of machine learning: linear algebra, calculus and optimization, and probability and statistics. All class files will be available here. For a more detailed outline of the course thus far, see the Course Skeleton.
- Lecture slides can be found by clicking on the lecture title for the appropriate day.
- All the materials and reading on the right column is optional, but reading (a subset of) these materials before each lecture might help digesting the content during lecture.
- Problem sets will be posted here, as well as their solutions.
This is a tentative schedule and is subject to change. Readings, slides, and assignments will be posted as the class goes on.
Optional readings. MML refers to Mathematics for Machine Learning by Marc Peter Deisenroth, A. Aldo Faisal, and Cheng Soon Ong. VMLS refers to Introduction to Applied Linear Algebra - Vectors, Matrices, and Least Squares by Stephen Boyd and Lieven Vandenberghe.
Story of the course. As the lectures go on, the goal will be to develop two main ideas from machine learning: least squares regression (LS) and gradient descent (GD). During each lecture, we will build these ideas with the mathematical tools from that lecture; at the same time, we’ll gradually develop a “picture” of LS and GD as the course goes on. An evolving 3D rendering of each “picture” will be linked in each module below.
Problem sets. The problem sets will usually look relatively long, but much of it is exposition – the problems in this course are mostly structured to guide you through the discovery or derivation of some result or perspective on a concept. As such, the problem sets serve the double purpose of some “required reading” interspersed with problems for you to fill in the gaps.
Lecture pace. It’s really easy, in my experience, to get lost in a math lecture when lots of derivations or proofs are involved. At the same time, though, it can often be intimidating to speak up for fear of asking a “dumb question” (no such thing!). To this end, during every lecture, I’ll have a fully anonymous interactive poll to keep an eye on how people are feeling during lecture and I’ll check it intermittently, especially during proofs. Access the poll on the Pacing page.
Unit reviews. At the end of each “pillar” of the course, we will hold an optional unit review session to make sure that everyone is on the same page before moving onto the next session. These will be informal recitations where we recap the Course Skeleton to get a big picture view and, more importantly, answer any questions and confusion you might have. The dates/times/locations will be posted here and on the Calendar.
Linear Algebra I (matrices, vectors, bases, and orthogonality)
- May 22
- PS 0 released (due May 30 11:59 PM ET) + Ed Announcement
- ps0_template.zip
- May 27
- Lecture: Vectors, matrices, and least squares
- MML 2.1 - 2.8, 3.1 - 3.3, VMLS 1.1-1.5, 2.1-2.3, 3.1-3.4, 5.1, 5.2, 6.1-6.4, 12.1-12.4, Regression (d=2)
- PS 1 released (due June 6 11:59 PM ET)
- ps1.pdf, ps1_student.zip, ps1.ipynb, ps1_tex.zip
- Reading Project released (due June 3 11:59 PM ET)
- project instructions
- May 29
- May 30
- DUE PS 0 due
- LS (Story thus far)
- Lecture 1.1: Least squares regression can be solved geometrically with the Pythagorean Theorem.
- Lecture 1.2: Least squares regression has a simpler solution with orthonormal bases.
- GD (Story thus far)
- Lecture 1.1, 1.2: Gradient descent with a “bowl-shaped” function gets us to the minimum.
Linear Algebra II (singular value decomposition and eigendecomposition)
- Jun 3
- Lecture: Singular Value Decomposition
- 3D SVD (unprojected), 3D SVD (u1, u2), 3D SVD (u1), Orthogonal Complement, MML 4.2, 4.4, 4.5, Daniel Hsu’s Computational Linear Algebra (CLA) course notes on SVD, Daniel Hsu’s CLA interactive example of “best-fitting 1d subspace”
- DUE Reading Project first evaluation due
- PS 2 released (due June 13 11:59 PM ET)
- ps2.pdf, ps2_student.zip, ps2.ipynb, ps2_tex.zip
- Jun 5
- Jun 6
- DUE PS 1 due
- LS (Story thus far)
- Lecture 2.1 & 2.2: The problem of least squares regression is unified under the pseudoinverse.
- GD (Story thus far)
- Lecture 2.1 (nothing new): Gradient descent with a “bowl-shaped” function gets us to the minimum.
- Lecture 2.2: On quadratic forms, it seems that gradient descent on three different types of shapes has different behavior: positive definite, positive definite, and indefinite.
Calculus and Optimization I (differentiation and Taylor Series)
- Jun 10
- Lecture: Differentiation and vector calculus
- “Peaks” Function, Derivative Ex. 1, Derivative Ex. 2, Derivative Ex. 3, MML 5.1 - 5.5, The Matrix Cookbook, Annotated Slides
- PS 3 released (due June 20 11:59 PM ET)
- ps3.pdf, ps3_student.zip, ps3.ipynb
- Jun 12
- Jun 13
- DUE PS 2 due
- LS (Story thus far)
- Lecture 3.1: We can derive the exact same OLS theorem from linear algebra section from just the tools of optimization and viewing the notion of least squares error as an “objective function.”
- GD (Story thus far)
- Lecture 3.1: We can now write down the algorithm for gradient descent. Intuitively, positive semidefinite or positive definite quadratic forms seem good for gradient descent.
- Lecture 3.2: Using Taylor’s theorem for the first-order approximation (linearization), we can provide intuition and a formal guarantee that gradient descent makes the function values decrease. The behavior of gradient descent depends on the learning rate eta: eta too big will result in erratic behavior but small enough eta results in stable convergence. This eta setting depends intimately on the second order information, or “smoothness” of the function
Calculus and Optimization II (optimization and convexity)
- Jun 17
- PS 4 released (due June 27 11:59 PM ET)
- ps4.pdf, ps4_student.zip, ps4.ipynb
- Jun 19
- Class rescheduled to Friday, June 20th due to Juneteenth
- Jun 20
- Lecture: Convexity and convex optimization (Changed time and location: 12:45pm - 4pm in CSB 451)
- MML 7.3, Convexity Definition in 3D, Convexity First-order Definition in 3D, Boyd and Vandenberghe’s Convex Optimization Chapters 1 - 3
- Jun 20
- DUE PS 3 due
- LS (Story thus far)
- Lecture 4.1: In some applications, it may be favorable to regularize the least squares objective by trading off minimizing the objective with the norm of the weights.
- Lecture 4.2: The least squares objective is a convex function (also: first-order definition); applying gradient descent takes us to a global minimum
- GD (Story thus far)
- Lecture 4.1: Nothing new here.
- Lecture 4.2: Applying gradient descent to beta-smooth, convex functions takes us to a global minimum. One such function is the least squares objective.
Probability and Statistics I (basic probability theory and statistical estimation)
- Jun 24
- Lecture: Basic Probability Theory, Models, and Data
- Regression setup w/ randomness, MML 6.1-6.4, Blitzstein and Hwang’s Ch. 9 on Conditional Expectation, Leo Breiman’s “Two Cultures” paper, Carlos Fernandez-Granda’s Probability for Data Science Overview
- PS 5 released (due July 4 11:59 PM ET), no programming part
- ps5.pdf, ps5_student.zip
- Final paper reading evaluation released. Evaluation due July 8 11:59 PM ET
- Jun 26
- Jun 27
- DUE PS 4 due
LS (Story thus far) :
GD (Story thus far) :
Probability and Statistics II (Maximum likelihood and Gaussian distribution)
- Jul 1
- Lecture: The Central Limit Theorem, “Named” Distributions, and MLE
- MML 6.1-6.8, MML Ch. 8, 3Blue1Brown’s video on the Central Limit Theorem
- Jul 3
- Lecture: Multivariate Gaussian
- 3Blue1Brown’s video on adding Gaussian distributions, 3Blue1Brown’s video on normalizing the Gaussian, MML Ch. 11 (Gaussian Mixture Models, not covered), OLS distribution with standard normal eps, true w = (1,1), MVN with mean (0, 0), Identity covariance, MVN with mean (0, 0), Diagonal covariance, MVN with mean (0, 0), Non-diagonal covariance, MVN with mean (1, 1), Non-diagonal covariance
- Jul 4
- DUE PS 5 due
- Jul 8
- DUE Final Project Evaluation due
- LS (Story thus far)
GD (Story thus far) :
Resources
I’ll update this with additional resources as the class progresses. Feel free to use these or ignore completely. If you know of any additional resources that you think would be helpful for the class, let me know and I’ll add it here!
LaTeX
- Overleaf, the Google Docs for LaTeX. Can be used for all the assignments in this class.
- Overleaf’s guide to learn LaTeX in 30 minutes
- David Xiao’s Beginner’s guide to LaTeX
- Eddie Kohler’s LaTeX usage notes. These might be worth a browse to rectify common stylistic problems with using LaTeX.
- Detexify, an applet to get the LaTeX command for any handwritten symbol.
In general, Googling an issue you’re having with LaTeX usually provides a plethora of solutions.
Python
- Whirlwind Tour of Python should have most everything you need to get up to speed with the programming required in this course.
- A condensed version of this Whirlwind Tour of Python can be found here: python_crashcourse.ipynb.
- Here is a video going through this crash course in case you want to get up to speed in video format.
Linear Algebra Prerequisites
If you need to refresh any linear algebra, these may be good resources.
- Linear Algebra and Applications by Gilbert Strang
- Gilbert Strang’s MIT Course on Linear Algebra
- Linear Algebra Done Wrong by Sergei Treil, available free as PDF here
- Daniel Hsu’s course notes for Computational Linear Algebra
- 3Blue1Brown’s Essence of Linear Algebra videos
Multivariable Calculus Prerequisites
If you need to refresh any multivariable calculus, these may be good resources.
- MIT OpenCourseware course on multivariable calculus
- Vector Calculus, Linear Algebra, and Differential Forms: A Unified Approach by Barbara Burke Hubbard and John H. Hubbard.
- Vector Calculus by Susan Jane Colley
Probability Theory and Statistics Prerequisites
If you need to refresh any probability and statistics, these may be good resources.
- Introduction to Probability for Data Science by Stanley H. Chan
- A First Course in Probability by Sheldon Ross.
- Introduction to Probability by Joseph K. Blitzstein and Jessica Hwang.
- Probability and Statistics for Engineers and Scientists by Ronald E. Wadpole.