Statistical machine learning at the University of Melbourne
We’ll put the lecture slides in the week that we cover the material, as well as pointers to the required reading. Reading references to Bishop relate to the book Pattern recognition and machine learning by Christopher M. Bishop, 2006. Reading references to Murphy relate to the book Machine Learning : A Probabilistic Perspective by Kevin Murphy, 2014; available in electronic ebook form using the link.
| Date | Topic | Materials |
| Tue 25/7 | Introduction; probability theory |
Slides:
01_intro_prob_theory.pdf
Reading: Bishop 1.1-1.2 |
| Thu 27/7 | Probabilistic models and parameter fitting |
Slides:
02_statistical_schools.pdf
Reading: Bishop 2.1*, 2.3*; 1.2.3-1.2.4 (* = first 2 pages) |
| Tue 1/8 | Linear regression; Intro to regularisation |
Slides:
03_linear_regression.pdf
Reading: Bishop 3.1.1, 3.1.2, 3.1.4 |
| Thu 3/8 | Logistic regression classifier; Basis expansion |
Slides:
04_logistic_regression.pdf
Reading: Bishop 4.3.2, 3.1* |
| Tue 8/8 | Iterative optimisation of loss functions; Model complexity and bias-variance analysis |
Slides:
05_optim_regularisation.pdf
Reading: Bishop 1.5.5, 3.2 |
| Thu 10/8 | Notes on vectors; Perceptron classifier |
Slides:
06_vectors_perceptron.pdf
Reading: Bishop 4.1.7 |
| Tue 15/8 | Artificial Neural Networks |
Slides:
07_backpropagation.pdf
Reading: Bishop 5.1, 5.2, 5.3 |
| Thu 17/8 | Artificial Neural Networks (cont.) |
Slides:
08_convnet_autoenc.pdf
Reading: Bishop 5.5.6 |
| Tue 22/8 | Support Vector Machines (hard margin) |
Slides:
09_hard_margin_svm.pdf
Reading: Bishop 7.1 |
| Thu 24/8 | Support Vector Machines (soft margin) |
Slides:
10_soft_margin_svm.pdf
Reading: Bishop 7.1.1 |
| Tue 29/8 | Kernel Methods |
Slides:
11_kernel_methods.pdf
Reading: Bishop 6.2 |
| Thu 31/8 | Ensemble Learning; Interim Summary |
Slides:
12_ensemble_revision.pdf
Reading: Bishop 14.2 |
| Tue 5/9 | Unsupervised Learning; Clustering |
Slides:
13_clustering_gmm.pdf
Reading: Bishop 9.1, 9.2 |
| Thu 7/9 | Expectation Maximisation Algorithm |
Slides:
14_em_algorithm.pdf
Reading: Bishop 9.4, 9.2.2 |
| Tue 12/9 | Principal Component Analysis; Multidimensional Scaling |
Slides:
15_dimred_pca_mds.pdf
Reading: Bishop 12.1, 12.4.3, Hastie 14.8 |
| Thu 14/9 | Manifold Learning; Spectral clustering |
Slides:
16_manifold_learning.pdf
Reading: Bishop 12.4.3, Hastie 14.5.3, 14.9 |
| Tue 19/9 | Bayesian inference |
Slides:
17_bayesian_inference.pdf
Reading: Bishop 2.3.6, 4.4-4.5 |
| Thu 21/9 | Bayesian inference (cont.) |
Slides:
18_bayesian_classification.pdf
Reading: Bishop 2.1, 3.3, 3.4 |
| Mon 25/9 | Non-teaching week | |
| Tue 3/10 | Probabilistic graphical models, fundamentals |
Slides:
20_pgm_basics.pdf
Reading: Bishop 8-8.3.1 |
| Thu 5/10 | Probabilistic graphical models, independence |
Slides:
21_pgm_indy.pdf
Reading: Bishop 8.2-8.3.2 |
| Tue 10/10 | Probabilistic graphical models, inference |
Slides:
22_pgm_elimination.pdf
Reading: Russell and Norvig Chapter 14 (see "Reading" on LMS) or Murphy 20.3 |
| Thu 12/10 | Probabilistic graphical models, statistical inference. Lecture cancelled, please listen to 2016 recording available under LMS in the reading/resources option. |
Slides:
23_pgm_stat_infer.pdf
Reading: Murphy 11.4.4 (generally dip into 11.3-11.4) |
| Tue 17/10 | Probabilistic graphical models, hidden Markov models and message passing |
Slides:
24_pgm_message_passing.pdf
Reading: Murphy 17 (mainly 17.4-17.5) |
| Thu 19/10 | Subject review |
Slides:
25_review.pdf
|