MIT 18.065 Matrix Methods in Data Analysis and Machine Learning: study plan and how long it takes
The MIT OpenCourseWare playlist is 36 videos and 28.2 hours long. Here's how long it takes at different paces, a week-by-week schedule built from the real videos, and a planner that turns it into your own daily plan.
36 videos · 28.2 hours · MIT OpenCourseWare
Gilbert Strang's MIT 18.065 (Spring 2018) applies linear algebra to data and machine learning: matrix factorizations, positive definite matrices, the SVD and low-rank approximation, norms, least squares, gradient descent and acceleration, the structure of neural networks, and Fourier and circulant matrices.
It's the natural follow-on to 18.06 for anyone heading into ML.
Who it's for: Learners who know linear algebra (ideally from 18.06) and want the matrix methods behind data science and deep learning.
How long does it take?
The full playlist is 28.2 hours across 36 videos. Here's roughly how many days that is at different daily watch times and speeds, counting video time only:
| Daily time | At 1x | At 1.5x | At 2x |
|---|---|---|---|
| 30 min a day | 57 days | 38 days | 29 days |
| 1 h a day | 29 days | 19 days | 15 days |
| 2 h a day | 15 days | 10 days | 8 days |
Make your own schedule
The playlist link is already filled in. Press Plan, set how much time you have each day and your speed, and you'll get a day-by-day schedule with a finish date and a calendar file. It reads the live playlist, so new videos are included.
Week-by-week schedule
A sample plan at 1 hour a day and normal speed, using whole videos so none is split across days. That's 5 weeks in total. For a different pace, start date or speed, use the planner above.
Week 1Videos 1–8 · 5h 13m
Day 1 · 15m
- 1Course Introduction of 18.065 by Professor Strang7m
- 2An Interview with Gilbert Strang on Teaching Matrix Methods in Data Analysis, Signal Processing,...8m
Day 2 · 52m
- 3Lecture 1: The Column Space of A Contains All Vectors Ax52m
Day 3 · 48m
- 4Lecture 2: Multiplying and Factoring Matrices48m
Day 4 · 49m
- 53. Orthonormal Columns in Q Give Q'Q = I49m
Day 5 · 49m
- 64. Eigenvalues and Eigenvectors49m
Day 6 · 45m
- 75. Positive Definite and Semidefinite Matrices45m
Day 7 · 54m
- 86. Singular Value Decomposition (SVD)54m
Week 2Videos 9–15 · 5h 48m
Day 8 · 47m
- 97. Eckart-Young: The Closest Rank k Matrix to A47m
Day 9 · 49m
- 10Lecture 8: Norms of Vectors and Matrices49m
Day 10 · 50m
- 119. Four Ways to Solve Least Squares Problems50m
Day 11 · 50m
- 12Lecture 10: Survey of Difficulties with Ax = b50m
Day 12 · 50m
- 13Lecture 11: Minimizing ‖x‖ Subject to Ax = b50m
Day 13 · 49m
- 1412. Computing Eigenvalues and Singular Values49m
Day 14 · 52m
- 15Lecture 13: Randomized Matrix Multiplication52m
Week 3Videos 16–22 · 5h 51m
Day 15 · 51m
- 1614. Low Rank Changes in A and Its Inverse51m
Day 16 · 51m
- 1715. Matrices A(t) Depending on t, Derivative = dA/dt51m
Day 17 · 43m
- 1816. Derivatives of Inverse and Singular Values43m
Day 18 · 51m
- 19Lecture 17: Rapidly Decreasing Singular Values51m
Day 19 · 49m
- 20Lecture 18: Counting Parameters in SVD, LU, QR, Saddle Points49m
Day 20 · 52m
- 2119. Saddle Points Continued, Maxmin Principle52m
Day 21 · 55m
- 2220. Definitions and Inequalities55m
Week 4Videos 23–29 · 6h 8m
Day 22 · 54m
- 23Lecture 21: Minimizing a Function Step by Step54m
Day 23 · 53m
- 2422. Gradient Descent: Downhill to a Minimum53m
Day 24 · 49m
- 2523. Accelerating Gradient Descent (Use Momentum)49m
Day 25 · 54m
- 2624. Linear Programming and Two-Person Games54m
Day 26 · 53m
- 2725. Stochastic Gradient Descent53m
Day 27 · 53m
- 2826. Structure of Neural Nets for Deep Learning53m
Day 28 · 53m
- 2927. Backpropagation: Find Partial Derivatives53m
Week 5Videos 30–36 · 5h 8m
Day 29 · 50m
- 30Lecture 30: Completing a Rank-One Matrix, Circulants!50m
Day 30 · 53m
- 3131. Eigenvectors of Circulant Matrices: Fourier Matrix53m
Day 31 · 47m
- 32Lecture 32: ImageNet is a Convolutional Neural Network (CNN), The Convolution Rule47m
Day 32 · 56m
- 3333. Neural Nets and the Learning Function56m
Day 33 · 29m
- 3434. Distance Matrices, Procrustes Problem29m
Day 34 · 35m
- 3535. Finding Clusters in Graphs35m
Day 35 · 38m
- 36Lecture 36: Alan Edelman and Julia Language38m
Tips for finishing
- Do 18.06 first; this course assumes it.
- The gradient descent and neural network lectures connect directly to deep learning courses. Watch them before CS231N or CS224N.