Note on nomenclature: This course is entitled as above for historical reasons.
However, today the term Machine Learning is more widely used to denote
the general body of statistical techniques for automatically detecting and modelling
patterns in data. Pattern Recognition may at times
refer to the use of a more conventional subset of these techniques, such as Neural Networks. This course should thus essentially be seen as an introduction to
Machine Learning.
[Machine Learning: A very quick introduction.]
[Introduction to a similar course, written by Amos Storkey at the University of Edinburgh.]
Serial no. | Topics | Lecture nos. | PRML Chapters | Other References | Slides |
1 | Introduction, Revision of Probability Theory and Distributions | 1–7 | 1, 2 | Review of Probability Theory; The Multivariate Gaussian Distribution | Very quick introduction; PRML Chapter 1 |
2 | Supervised Learning: Linear Regression Models | 8–9 | 3 | Regression notes | PRML Chapter 3 |
3 | Supervised Learning: Classification, Linear Discriminant Analysis | 10–11 | 4 | Generative learning notes | Linear models for classification |
4 | Supervised Learning: Kernels, Support Vector Machines | 12–16 | 6, 7 | SVM notes; SVM tutorial; Properties of kernels; Platt's SMO paper; SVM demo | |
5 | Supervised Learning: Neural Networks, Deep Learning | 17–21 | 5 | ANN tutorial; Geoff Hinton lecture on deep learning; Convolutional neural nets: tutorial, demos; Sparse autoencoder notes; Primate visual system paper | Google's unsupervised deep learning |
6 | Feature Selection | 22 | |||
7 | Naïve Bayes, Probabilistic Graphical Models | 23–25; 29 | 8 | Naïve Bayes notes; MRF notes; Bayesian Networks notes | Graphical Models |
8 | Unsupervised Learning: Clustering, Mixture Models, Expectation-Maximisation | 26–28 | 9 | K-means notes; GMM notes; EM notes | Clustering |
9 | Unsupervised Learning: Latent Variables, Component Analysis | 30–34 | 12 | PCA notes; PCA tutorial; Factor analysis notes; ICA notes; ICA overview; Le et al. paper on reconstruction ICA | PCA; ICA for feature learning |
10 | Hidden Markov Models | 35–38 | 13 | HMM notes; Microsoft's photo-real talking head [demo paper] | |
11 | Semi-Supervised Learning | 39–40 | SSL literature survey | ||
12 | Computational Learning Theory | 41 | Jayadeva's MCM paper |