Undergraduate Teaching

Engineering Tripos Part IIB, 4F10: Statistical Pattern Processing, 2016-17

Engineering Tripos Part IIB, 4F10: Statistical Pattern Processing, 2016-17

Not logged in. More information may be available... Login via Raven / direct.

PDF versionPDF version

Leader

Prof M Gales

Lecturer

Prof M Gales

Timing and Structure

Michaelmas term. 14 lectures + 2 examples classes. Assessment: 100% exam

Prerequisites

Part IIA Modules 3F1 and 3F3 advisable

Aims

The aims of the course are to:

  • describe the basic concepts of statistical pattern processing and some of the current techniques used in pattern classification.

Objectives

As specific objectives, by the end of the course students should be able to:

  • understand the basic principles of pattern classification.
  • understand Expectation-Maximisation as a general optimisation technique.
  • understand current classification schemes such as Support Vector Machines and Gaussian Processes.
  • apply pattern processing techniques to practical applications.

Syllabus

Introduction (1L)

Statistical pattern proecessing, Bayesian decision theory, generalisation. 

Multivariate Gaussian Distributions and Decision Boundaries (1L)

Multivariate Gaussian PDFs, maximum likelihood estimation, decision boundaries, classification cost, ROC curves. 
 

Gaussian Mixture Models (1L)

Mixture models, parameter estimation, EM for discrete latent variables.
 

Expectation Maximisation (1L)

Latent variables both continuous and discrete, proof of EM, factor analysis.
 

Mixture and Product of Experts (1L)

Combining multiple classifiers/predictors, gating functions, products versus mixtures.
 

Resticted Boltzman Machines (1L)

Structure of restricted Boltzman machines, contrastive divergence.
 

Linear Classifiers (1L)

Single layer perceptron, perceptron learning algorithm, Fisher's linear discriminant analysis, limitations.
 

Multi-Layer Perceptrons (2L)

Basic structure, posterior modelling, regression, error back propogation, learning rates, second order optimisation methods, "deep" topologies, network initialisation. 
 

Support Vector Machines (2L)

Maximum margin classifiers, handling non-separable data, training SVMs, non-linear SVMs, kernel functions.
 

Classification and Regression Trees (1L)

Decision trees, query selection, multivariate decision trees.
 

Non-Parametric Techniques (1L)

Parzen windows, K-nearest neighbours, nearest neighbour rule.
 

Speaker Recognition and Verification (1L)

Speaker recognition/verification task, GMMs and MAP adaptation, SVM-based verification.
 

Booklists

Please see the Booklist for Group F Courses for references for this module.

Assessment

Please refer to Form & conduct of the examinations.

UK-SPEC

The UK Standard for Professional Engineering Competence (UK-SPEC) describes the requirements that have to be met in order to become a Chartered Engineer, and gives examples of ways of doing this.

UK-SPEC is published by the Engineering Council on behalf of the UK engineering profession. The standard has been developed, and is regularly updated, by panels representing professional engineering institutions, employers and engineering educators. Of particular relevance here is the 'Accreditation of Higher Education Programmes' (AHEP) document which sets out the standard for degree accreditation.

The Output Standards Matrices indicate where each of the Output Criteria as specified in the AHEP 3rd edition document is addressed within the Engineering and Manufacturing Engineering Triposes.

Last modified: 31/05/2016 09:13