732A96 Advanced Machine Learning
Course descriptionThe course covers some advanced models in machine learning. The models are analyzed mainly from a Bayesian perspective.
The course is organized into four topics:
- Graphical Models
- Hidden Markov Models
- Reinforcement Learning
- Gaussian Process Regression and Classification
Each topic includes lectures, a computer lab and a follow-up seminar.
Course literature- Chapters from the book Pattern Recognition and Machine Learning by Bishop.
- Chapters from the book Probabilistic Graphical Models: Principles and Techniques by Koller and Friedman.
- The article A Review of Bayesian Networks and Structure Learning by Koski and Noble.
- Chapters from the book Causality: Models, Reasoning, and Inference by Pearl.
- The article An Introduction to Hidden Markov Models and Bayesian Networks by Ghahramani.
- The article Explicit-Duration Markov Switching Models by Chiappa.
- Chapters from the book Reinforcement Learning: An Introduction by Sutton and Barto.
- The article Playing Atari with Deep Reinforcement Learning by Mnih et al.
- Chapters from the book Gaussian Processes for Machine Learning by Rasmussen and Williams.
Topic 1: Graphical ModelsTeacher: Jose M. Peña
Read: Bishop chapter 8. Koller and Friedman chapter 9.2-9.4. Lauritzen and Spiegelhalter. Pearl chapters 1-3.
Lecture 1: Bayesian and Markov Networks Slides
Lecture 2: Probabilistic Inference Slides
Lecture 3: Parameter Learning Slides
Lecture 4: Structure Learning Slides
Lecture 5: Causal Inference Slides
Lab 1: Lab. Submit your report via LISAM. Note that you have to submit an individual and a group report. A group consists of four people. The group is created when you submit.
Topic 2: Hidden Markov ModelsTeacher: Jose M. Peña
Read: Bishop chapter 13.1-13.2. Ghahramani. Chiappa sections 3.1-3.3.
Lecture 6: Dynamic Bayesian Networks and Hidden Markov Models Slides
Lecture 7: Autoregressive and Explicit-Duration Hidden Markov Models Slides
Lab 2: Lab. Submit your report via LISAM. Note that you have to submit an individual and a group report.
Topic 3: Reinforcement LearningTeacher: Jose M. Peña
Read: Sutton and Barto chapters 1-7, 13 and 16. Mnih et al.
Lecture 8: Q-Learning Algorithm Slides
Lecture 9: REINFORCE and Deep Q-Learning Algorithms Slides
Lab 3: Lab. Submit your report via LISAM. Note that you have to submit an individual and a group report.
Code: RL_Lab1.R, RL_Lab2_Colab.ipynb
Topic 4: Gaussian Process Regression and ClassificationTeacher: Jose M. Peña
Read: Bishop chapters 6.4.1-6.4.6. Rasmussen and Williams chapters 2.1-2.5, 3.1-3.4.1, 3.7, 5.1-5.4.1.
Lecture 10: Gaussian Process Regression Slides
Lecture 11: Kernels, Hyperparameter Learning and More Slides
Lecture 12: Gaussian Process Classification Slides
Lab 4: Lab. Submit your report via LISAM. Note that you have to submit an individual and a group report.
Extras: Web applet for playing around with a simple GP | Chapter on kernels from David Duvenaud's PhD thesis.
Code: R code for simulating from a GP | Quick demo of the R package kernlab.
Data: Lidar | CanadianWages
Page responsible: Jose M. Pe˝a
Last updated: 2021-09-20