1. To provide a theoretical understanding of modern deep learning.
2. To provide a critical understanding of the practical application of deep learning in the modern workplace.
3. To provide a deep insight into the usage of current deep learning libraries.

(M1) A comprehensive understanding of the nature of deep learning in the context of modern computing capabilities.

(M2) A systematic understanding of mathematical foundations and algorithmic principles of deep learning.

(M3) A critical understanding of the process of deploying deep learning systems and the limitations involved.

(M4) A practical ability to apply the techniques of deep learning using current deep learning libraries.

(S1) Communication skills in electronic as well as written form.

(S2) Self-direction and originality in tackling and solving problems.

(S3) An ability to act autonomously and professionally when planning and implementing solutions to computer science problems.

(S4) Experience of working in development teams, respecting others, co-operating, negotiating/persuading, awareness of interdependence with others.

Week 1: Foundations.
Introduction to deep learning, historical context, application context, available libraries and tool kits, practical aspects of Deep Learning, formulating a deep learning problem, case studies.

Week 2: Cloud computing.
The theory and operation of cloud computing, cloud security, introduction to a common cloud platform to support deep learning, setting up a cloud-based deep learning application.

Week 3: Neural Networks (NN):
Background, perceptrons, multi-layer neural networks, propagation, designing a neural network architectures.

Week 4: Regularization and optimizations.
Hyperparameter tuning, batch normalization, regularisation, optimization algorithms.

Week 5: Convolutional Neural Networks (CNNs)
Theoretical underpinning, convolution, pooling, applications of CNNs in image analysis, implementing a CNN using a popular deep learning library.

Week 6: Recurrent Neural Networks (RNNs)
Theoretical underpinnin g, applications of RNNs of language modelling, implementing a CNN using a popular deep learning library.

Week 7: Generative Adversarial Networks (GANs)
Theoretical underpinning, generators and discriminators, loss functions and optimizers, implementing a GAN using a popular deep learning library.

Week 8: Deep Reinforcement Learning.
Comparison with other forms of machine learning, the reward concept, the reinforcement learning process, deep learning for reinforcement learning, implementing reinforecment learning using a popular deep learning library.

The mode of delivery is by online learning, facilitated by a Virtual Learning Environment (VLE). This mode of study enables students to pursue modules via home study while continuing in employment. Module delivery involves the establishment of a virtual classroom in which a relatively small group of students (usually 10-25) work under the direction of a faculty member. Module delivery proceeds via a series of eight one-week online sessions, each of which comprises an online lecture, supported by other eLearning activities, posted electronically to a public folder in the virtual classroom. The mode of learning includes a range of required and optional eLearning activities, including but not limited to: lecture casts, live seminars, self-assessment opportunities, and required and suggested further reading and try-for-yourself activities. Communication within the virtual classroom is asynchronous, preserving the requirement that students are able to pursue the module in their own time, within the weekly time-frame of each online session. An important element of the module provision is active learning through collaborative, cohort-based, learning using discussion fora where the students engage in assessed discussions facilitated by the faculty member responsible for the module. This in turn encourages both confidence and global citizenship (given the international nature of the online student body).