ECE544NA: Pattern Recognition (Fall 2019)
Course Information
ECE 544NA is a special topics course in pattern recognition, and content varies every year. In Fall 2019, the course will cover three main areas, (1) disciminative models, (2) generative models, and (3) reinforcement learning. See course syllabus for more details.The goal of the course is to provide an understanding of recent research topics in pattern recognition. After having completed the class, students should be familiar with the underlying theory and software that is frequently used in publications related to pattern recognition.
Pre-requisites: Probability, linear algebra, and proficiency in Python, MATLAB or equivalent.
Recommended Text: (1) Deep Learning by Ian Goodfellow and Yoshua Bengio and Aaron Courville, (2) Pattern Recognition and Machine Learning by Christopher Bishop, (3) Graphical Models by Nir Friedman and Daphne Koller and (4) Machine Learning: A Probabilistic Perspective by Kevin Murphy.
Grading: Scribing/project 1/3, midterm 1/3 and final exam 1/3 of the grade. Weights might get adjusted during the course.
Late Policy: No late submission will be accepted after the due date.
Instructor & TAs
Alexander Schwing
InstructorEmail: aschwing[at]illinois.edu
Office Hour: Tues. 13:00-14:00PM
Room: CSL 103
Website: [link]
Safa Messaoud
Teaching AssistantEmail: messaou2[at]illinois.edu
Office Hour: Friday 9:00AM-10:00AM
Room: ECEB 4034
Raymond A. Yeh
Teaching AssistantEmail: yeh17[at]illinois.edu
Office Hour: Friday 9:00AM-10:00AM
Room: ECEB 4034
Website: [link]
Lectures
The syllabus is subject to change.
Event | Date | Description | Materials |
---|---|---|---|
Lecture 1 | Aug. 27 | Intro/Nearest Neighbor | [Slides] [Rec] [Audio] |
Lecture 2 | Aug. 29 | Linear Regression | [Slides] [Rec] [Audio] |
Lecture 3 | Sep. 3 | Logistic Regression | [Slides] [Rec] [Audio] |
Lecture 4 | Sep. 5 | Optimization Primal | [Slides] [Rec] [Audio] |
Lecture 5 | Sep. 10 | Optimization Dual | [Slides] [Rec] [Audio] |
Lecture 6 | Sep. 12 | SVM | [Slides] [Rec] [Audio] |
Lecture 7 | Sep. 17 | Multiclass Classification and Kernel Methods | [Slides] [Rec] [Audio] |
Lecture 8 | Sep. 19 | Deep Nets 1 (Layers) | [Slides] [Rec] [Audio] |
Lecture 9 | Sep. 24 | Deep Nets 2 (Backprop) & Pytorch 1 | [Slides] [Rec] [Audio] |
Lecture 10 | Sep. 26 | Pytorch 2 | [Slides] [Rec] [Audio] |
Lecture 11 | Oct. 1 | Review | [Slides] [Rec] [Audio] |
Midterm | Oct. 3 | Midterm | [Slides] [Rec] [Audio] |
Lecture 13 | Oct. 8 | Ensemble Methods (Boosting/Random Forest/Deep Nets) & Regularization/Cross-Validation | [Slides] [Rec] [Audio] |
Lecture 14 | Oct. 10 | Structured Prediction (exhaustive search, dynamic programming) | [Slides] [Rec] [Audio] |
Lecture 15 | Oct. 15 | Structured Prediction (ILP, LP relaxation) | [Slides] [Rec] [Audio] |
Lecture 16 | Oct. 17 | Learning in Structured Models | [Slides] [Rec] [Audio] |
Lecture 17 | Oct. 22 | Learning Theory | [Slides] [Rec] [Audio] |
Lecture 18 | Oct. 24 | PCA and SVD | [Slides] [Rec] [Audio] |
Lecture 19 | Oct. 29 | k-Means | [Slides] [Rec] [Audio] |
Lecture 20 | Oct. 31 | Gaussian Mixture Models | [Slides] [Rec] [Audio] |
Lecture 21 | Nov. 5 | Expectation Maximization | [Slides] [Rec] [Audio] |
Lecture 22 | Nov. 7 | Hidden Markov Models | [Slides] [Rec] [Audio] |
Lecture 23 | Nov. 12 | Variational Auto-Encoders | [Slides] [Rec] [Audio] |
Lecture 24 | Nov. 14 | Generative Adversarial Nets | [Slides] [Rec] [Audio] |
Lecture 25 | Nov. 19 | Autoregressive Methods, Recurrent Neural Nets, Graph Convolutions | [Slides] [Rec] [Audio] |
Lecture 26 | Nov. 21 | MDP | [Slides] [Rec] [Audio] |
No Lecture | Nov. 26 | Fall Break | [Slides] |
No Lecture | Nov. 28 | Fall Break | [Slides] |
Lecture 27 | Dec. 3 | Q-Learning | [Slides] [Rec] [Audio] |
Lecture 28 | Dec. 5 | Policy Gradient | [Slides] [Rec] [Audio] |
Lecture 29 | Dec. 10 | Review | [Slides] [Rec] [Audio] |
Final Exam | Dec. 13 | Final Exam, Rooms 2015/3017 | [Slides] |