Module Details
Module Code: |
SYST C4606 |
Module Title:
|
Deep Learning
|
Title:
|
Deep Learning
|
Module Level:: |
8 |
Module Coordinator: |
Frances Hardiman
|
Module Author:: |
James Garland
|
Module Description: |
Deep neural networks can inform both the contents of an image or video frame and the content's location within the image boundaries. Additionally, neural networks can manipulate images and video frames. This module investigates methods of image classification, location, and manipulation. The module also examines optimisation of the computation and storage of these neural network models' immense data to provide the student with a demonstrable understanding of the advanced neural network features.
|
Learning Outcomes |
On successful completion of this module the learner will be able to: |
# |
Learning Outcome Description |
LO1 |
Design AI modules that identify features in images. |
LO2 |
Develop AI modules to track the movement of features in images. |
LO3 |
Manage image manipulation within image sets, e.g., using GANs. |
LO4 |
Improve the performance of the neural network model. |
LO5 |
Complete a project as an individual or in a small group to design and implement a solution for a real world problem. |
Dependencies |
Module Recommendations
This is prior learning (or a practical skill) that is recommended before enrolment in this module.
|
9271 |
COMP C4602 |
Computer Vision |
9655 |
ELEC C4602 |
Artificial Intelligence and Machine Learning |
Co-requisite Modules
|
No Co-requisite modules listed |
Additional Requisite Information
|
No Co Requisites listed
|
Indicative Content |
Image classification and localisation
Models to find the best classification accuracy and localisation of images. Localisation of objects in an image or video stream.
|
Semantic segmentation
Location and movement of items within a frame.
|
Image Manipulation
Generative variational auto-encoders, generative adversarial Networks (GANs), spectral normalisation.
|
Optimisation
Optimisation techniques such as pruning, activation functions, compression, and alternative number representation.
|
Ethics, Safety, and Trustworthiness
Algorithm and data bias, model safety and EU trustworthiness policy, GDPR considerations.
|
Module Content & Assessment
|
Assessment Breakdown | % |
Continuous Assessment | 20.00% |
Project | 40.00% |
Practical | 40.00% |
AssessmentsFull Time
No End of Module Formal Examination |
Reassessment Requirement |
Repeat examination
Reassessment of this module will consist of a repeat examination. It is possible that there will also be a requirement to be reassessed in a coursework element.
|
SETU Carlow Campus reserves the right to alter the nature and timings of assessment
Module Workload
Workload: Full Time |
Workload Type |
Workload Category |
Contact Type |
Workload Description |
Frequency |
Average Weekly Learner Workload |
Hours |
Lecture |
|
Contact |
No Description |
Every Week |
2.00 |
2 |
Laboratory |
|
Contact |
No Description |
Every Week |
3.00 |
3 |
Independent Learning |
|
Non Contact |
No Description |
Every Week |
6.00 |
6 |
Total Weekly Contact Hours |
5.00 |
Module Resources
|
Recommended Book Resources |
---|
-
GĂ©ron, A.. (2019), Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow: Concepts, tools, and techniques to build intelligent systems., O'Reilly Media.
-
Goodfellow, I., Bengio, Y., Courville, A. (2016), Deep learning, 2. Cambridge: MIT press.
-
Bishop, C. (2006), Pattern Recognition and Machine Learning, Springer.
| Recommended Article/Paper Resources |
---|
-
Dufresne-Camaro, C. O., Chevalier, F.,
& Ahmed, S. I.. (2020), Computer vision applications and their
ethical risks in the global south., Canadian Human-Computer Communications
Society (CHCCS), p.10,
| Other Resources |
---|
-
Kaggle. (2021), Kaggle, Kaggle Inc.,
-
Stanford Image Lab. (2020), ILSVRC, Stanford Image Lab,
-
EU. (2021), Ethics guidelines for trustworthy AI, EU, EU,
|
|