We link here handouts and lecture videos for ECE 283, a graduate course on machine learning created by Prof. Madhow. The goal is to employ a systematic framework grounded in probabilistic reasoning and optimization, in order to gain a fundamental understanding of the “best” approaches out there for topics such as unsupervised and supervised learning, sparsity-centric techniques, and Monte Carlo techniques. Insights into algorithms are developed via software-intensive homework assignments employing toy data sets in which the “right answer” is known, while real data is explored via class projects.
- 1. Introduction Handout
- 2. Model-based classification and Gaussian distributions Handout Videos
- 3. Logistic Regression Handout Videos
- Homework 1 (Spring 2020): Software lab on logistic regression (simulated data)
- 4. Neural Networks Handout Videos
- Homework 2 (Spring 2020): Software lab on neural networks (simulated data)
- 5. K Means, Gaussian Mixtures, and the EM Algorithm (Unsupervised Learning I) Handout Videos
- Homework 3 (Spring 2020): Software lab on K-means and Gaussian mixture modeling via EM algorithm
- 6. Unsupervised Learning with DNNs: VAEs and GANs Handout Videos
- 7. Dimension Reduction and Sparsity-Based Techniques Videos
- 7.1. Curse of Dimensionality and the Motivation for Sparsity Handout
- 7.2. SVD: Geometry and Application to PCA Handout
- 7.3. Linear Regression and Sparse Linear Regression Handout
- 7.4. Random Projections: Dimensionality Reduction and Compressive Sensing Handout
- 7.5. More Applications of Sparsity (Sparse coding & dictionary learning; Matrix completion) Handout
- Homework 4 (Spring 2020): Software lab on sparse modeling (PCA, compressive sensing and sketching)
- 8. Monte Carlo Techniques Handout
Lecture Videos are hosted on the following Youtube Channel, with playlists associated with each chapter: Machine Learning: A Signal Processing Perspective