DEEP LEARNING AND ADVANCED AI - 2017/8

Module code: COM3025

Module provider

Computer Science

Module Leader

CHEONG TOOK C Dr (Computer Sci)

Number of Credits

15

ECT Credits

7.5

Framework

FHEQ Level 6

JACs code

I270

Module cap (Maximum number of students)

N/A

Module Availability

Semester 2

Overall student workload

Assessment pattern

Assessment type Unit of assessment Weighting
Practical based assessment PRACTICAL ASSIGNMENTS ON KEY TOPICS TO BE COMPLETED IN THE FIRST 8 WEEKS 20
Coursework DEVELOPMENT OF NEW APPLICATIONS/FUNCTIONS BASED ON THE ALGORITHMS DISCUSSED IN THE MODULE 40
Examination 2 HOUR CLOSED BOOK UNSEEN EXAMINATION 40

Alternative Assessment

N/A

Prerequisites / Co-requisites

COM2028 Introduction to Artificial Intelligence; background in Linear algebra and probability will be helpful.

Module overview

Deep learning has shown its success in many areas including computer vision, speech and audio processing, natural language processing, robotics, bioinformatics and chemistry, video games, search engines, online advertising and finance. Deep learning is a particular kind of machine learning technique that allows computer systems to improve with experience and data, and achieves great power and flexibility by learning to represent the world as a nested hierarchy of concepts. In recent years, deep learning has seen tremendous growth in its popularity and usefulness, due in large part to more powerful computers, larger datasets and techniques to train deeper networks.

This module introduces a wide range of deep learning and other state of art techniques in AI for solving real world problems. Basic concepts on statistics and applied maths that thread through key elements in machine learning techniques will be discussed throughout the module. Students will study how to build suitable AI systems that can operate in complicated, real-world environments. The module also prepares students to explore further challenges and opportunities to improve deep learning and AI and bring them to new frontiers.

Module aims

Learning outcomes

Attributes Developed

Attributes Developed

C - Cognitive/analytical

K - Subject knowledge

T - Transferable skills

P - Professional/Practical skills

Module content



Vectors and Matrices



Linear Algebra Basics


Data Representation


Non-Negative Matrix Factorization


Application Examples: Collaborative Filtering, Finding Similar Users, Recommending Items




Data Visualisation



Principle component analysis (PCA)


Mutidimensional Scaling


t-Distributed Stochastic Neighbour Embedding (t-SNE)


Application Examples: Visualising High Dimensional Data




Machine Learning Basics



Learning Algorithms


Overfitting and Underfitting


Hyperparameter and Cross Validation


Linear Regression and Stochastic Gradient Decent




Convolutional Neural Network (CNN)



CNN architecture


Convolution Layer, Pooling and Fully Connected Layers


Activation Functions


Application Examples: Object Recognition in Visual Data, Search Engine




Recurrent neural networks (RNN)



Deep RNN


Long Short Term Memory (LSTM)


Application examples: Generating Text, Image Captioning, Speech Recognition, sentiment analysis




Autoencoder and Restricted Boltzmann Machines



Denoising Autoencoder


Deep Boltzmann Machines and Deep Belief Networks


Application Examples: Dimensionality Reduction, Classification, Collaborative Filtering, 




Reinforcement Learning (RL)



Key Elements in Reinforcement Learning


Deep Reinforcement Learning


Application Examples: Stunt Manoeuvres in a Helicopter, Play Chess, Manage an Investment Portfolio, Make a humanoid robot walk, Play many different Atari games better than humans





Current Research

Methods of Teaching / Learning

The module will develop analytical skills and the understanding of the subject area through:

• Lectures

• tutorials

• In-class discussion

The module will develop practical skills through:

• Lab sessions

• Coursework

All activities will be co-ordinated via the module webpage on the Surrey Learn.

Assessment Strategy

The assessment strategy is designed to provide students with the opportunity to demonstrate that they have achieved the module learning outcomes.

 

Thus, the summative assessment for this module consists of:



Completion of lab assignments.

Deadlines: All lab assignments are due in week 8 or during the Easter break. However, students are encouraged to submit individual assignment regularly to SurrreyLearn in order to receive timely feedback from the lecturer. A diagnostic test based on all lab assignments will be available on SurreyLearn in week8. Students will have a minimum 10 days to complete the test.  See the detailed dates on SurreyLearn.


Coursework project: deliverables will be in the form of a poster and the demonstration of the developed work. Due: week 10 - week 12.


Unseen exam



 

Formative assessment and feedback:

Between week 1-7, students will be guided to work on lab exercises and are encouraged to submit their work regularly through SurreyLearn in order to receive individual feedback. Students will be able to complete the coursework project successfully once the foundation of the coursework is built through lab exercises. The discussion on coursework will start from week 5 leading up to its completion date. Individual Feedback on the coursework project will be given as soon as possible within two weeks or before the exam whichever is sooner. Students will be able to gauge their progress through these feedbacks at different stages.

 

Reading list

Reading list for DEEP LEARNING AND ADVANCED AI : http://aspire.surrey.ac.uk/modules/com3025

Programmes this module appears in

Programme Semester Classification Qualifying conditions
Computer Science BSc (Hons) 2 Optional A weighted aggregate mark of 40% is required to pass the module

Please note that the information detailed within this record is accurate at the time of publishing and may be subject to change. This record contains information for the most up to date version of the programme / module for the 2017/8 academic year.