14. février 2019 - 17:30
Partager sur:

Foundations of Deep Learning, evening course | | jeudi, 14. février 2019

Course aims
This evening course, taught in person over 6 weekly sessions, aims to give developers/technical personnel a solid grounding in the field of deep learning, assuming no prior knowledge of the field. Regular practical sessions are combined with lecture-style content to consolidate the course material. By the end of the course you will understand the concepts underlying neural networks and deep learning, how these models are designed and trained, and gain an intuition of the knowledge encoded in these networks. You will also learn the basics of the Tensorflow framework and be able to design, build, train and apply your own network models. 
Summary of syllabus

Introduction to artificial intelligence (AI), machine learning (ML) and deep learning (DL). Basic important concepts underpinning ML and DL.

Mathematical prerequisites review.

Fundamentals of neural networks: distributed representations, concept hierarchy. Layer transformations, loss functions, network training and optimisers.

Feedforward networks / multilayer perceptrons.

Network design and initialisation: Xavier initialisation, batch normalisation, dropout layers, weight and layer normalisation.

Convolutional neural networks. Convolutional and pooling operations, strides, padding, transposed convolutions.

Analysis, interpretation and manipulation of trained neural networks.

Practical sessions to practise and consolidate course material. These sessions comprise an introduction to the Tensorflow deep learning framework to implement many of the concepts covered.

This course will be primarily led by Dr Kevin Webster, Honorary Research Fellow in Mathematics at Imperial College London. Kevin recently completed teaching the graduate level course on Deep Learning in the mathematics department at Imperial College London in Autumn 2018.
Who is this course for? 
This course presumes pre-existing technical proficiency. You might be a web developer, backend developer or data scientist interested in expanding your skills to include deep learning.
It will be assumed you have:

Basic knowledge of Git, Virtual Machines, Code Editors and Terminal Commands.

Basic knowledge of Object Oriented Programming.

Basic knowledge of Web development tools and environments.

Your own laptop to use during the session (inc. min. 60Gb of free space in HD, min. 8Gb of RAM (DDR3), min. 2.9GHz Dual Core i5, WiFi and/or Ethernet Port, full admin rights for installing and running tools and applications, Tensorflow pre-installed).

Understanding of core mathematical concepts (including linear algebra, probability theory, standard statistical distributions, information theory, elementary calculus). These topics will be briefly refreshed but not taught from scratch. If you have not previously learnt these topics, we suggest you take the "Foundations of Mathematics for Machine Learning" course first.


17.30-18.00: Registration
18.00-19.15: Session 1   
19.15-19.30: Break
19.30-20.45: Session 2
20.45-21.00: Wrap up

This course will take place every Thursday evening for 6 weeks, starting on Thursday 14th February 2019.

The course will take place in Central London, UK with easy access to a main tube station - the exact venue will be confirmed in the near future. 
If you have any questions about the course, please email lydia@