Home | Continuing Professional Development Programme



Introduction to Machine Learning

27th February – 3rd March 2017

Objectives

Over the last few years the field of machine learning has become very active mainly because of two major innovations, viz the invention of deep learning algorithms and the development of echosystems which made users able to program parallel computational platforms with greater ease. Currently people are trying to use machine learning in a range of uses starting from space data analysis to farming (e.g. ). New startups are venturing into interesting uses of machine learning and established industries are investigating how they can benefit as well. This short course is intended to expose the audience to the basics of machine learning and to get them started with using one of the powerful tools available namely Google TensorFlow.

|Exit level outcomes |Specific learning outcomes |

|Understand the basics of statistical machine learning |Can identify the data required for a particular machine learning |

| |problem |

|Appreciate the working of a single artificial neuron and a basic | |

|neural network | |

|Code a basic neural netowrok in Python |Apply a simple single layered artificail neural network to solve a|

| |simple data classification problem |

|Get started with TensorFlow |Use tensorflow to run some basic examples |

|Use tensorflow to code multilayer neural netowork |Apply tensorflow to solve a complicated pattern classification |

| |problem. |

Course Content

|Topics (including lab sessions) |Contact Hours |

|Introduction to machine learning and pattern recognition |4-5 |

|Bayes rule | |

|Statistics and linear algebra | |

|Parametric Classifiers |7-8 |

|Likelihood Ratio | |

|Discriminant Functions and Surfaces | |

|White Covariance Matrix and Its Implications | |

|Linear Classifiers | |

|Quadratic Classifiers | |

|MLE Classifiers | |

|Error Estimation for Classification Algorithms | |

|Huges Phenomena | |

|Non-parametric Classifiers |4-5 |

|Parzen Window Classifiers | |

|k-NN Classifiers | |

|Fuzzy Classifiers | |

|Unsupervised Classifiers |5-6 |

|Parametric Clustering | |

|Nonparametric Clustering | |

|ANN & Support Vector Machines |7-8 |

|Perceptron | |

|Structure of Artificial Neural Network | |

|Error Back-propagation and network regularization | |

|Kernel Methods | |

|Support Vector Machines | |

|Deep Learning |7-8 |

|Convolutional Neural Networks | |

|Recursive Neural Networks | |

|Deep Auto-encoders | |

|Machine Learning Strategies |4-5 |

|No Free-lunch Rule | |

|Re-sampling for Classifiers Testing | |

|Classifiers Design and Validation | |

Course Convenor

A/Prof. Amit Kumar Mishra has been working in the field of statistical signal processing and radar system development for past 12 years. He is an Associate Professor with the Department of Electrical Engineering, University of Cape Town. He is a Senior Member of IEEE and has more than 25 journal papers in ISI listed journals and is an inventor/co-inventor in eight patent applications.

A/Prof. Mishra shall be assisted by Mr. Jarryd Son who is a research scholar at University of Cape Town working in the domain of cognitive robotics.

Course Information

Who should attend?

Working engineers and software developers interested in the emerging field of machine learning and to gain some hands on using SciKit Learn and Google's TensorFlow toolboxes.

Format

The course is intensive and will take place over five days, and consists of lectures as well as simulation based lab modules. It is highly advised that the attendees come with their own laptop (at least one per group of two).

Cost

Standard registration R11 000.00

UCT Staff and Students R5 500.00

Students from other tertiary Inst R8 250.00

The course fee includes online course notes as well as lunch and refreshments.

Payment information will be sent on receipt of an application form.

Certificates

A certificate of attendance will be awarded to all course members who attend a minimum of 80% of the lectures.

CPD Credits

The course is registered with the Engineering Council of South Africa and is accredited for the award of CPD points, which are required for continuing professional registration. The ECSA course code is UCTIML17

Applications and cancellations

Registration forms are available on the website: cpd.uct.ac.za/cpd/cpdcourses/2017

Closing date for applications is one week before the start of the course

Payment is due one week before the start of a course.

Registration enquiries: Heidi Tait or Sandra Jemaar at ebe-cpd@uct.ac.za or tel: 021 6505793

Technical enquiries: A/Prof Amit Mishra at amit.mishra@uct.ac.za

Cancellations must be received one week before the start of the course, or the full course fee will be charged.

Further information regarding venue, directions and daily contact times will be forwarded to course delegates in the week prior to the course.

-----------------------

Continuing Professional Development Programme

Faculty of Engineering & the Built Environment

University of Cape Town

New Engineering Building, Upper Campus, University of Cape Town

Private Bag X3, Rondebosch, 7701

Tel: ++27 (0)21 6505793; Fax: ++27 (0)21 6503082; email: ebe-cpd@uct.ac.za; web: cpd.uct.ac.za

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download