Module Code - Title:

EE6008 - DEEP LEARNING AT THE EDGE

Year Last Offered:

2023/4

Hours Per Week:

Lecture

2

Lab

2

Tutorial

0

Other

0

Private

6

Credits

6

Grading Type:

N

Prerequisite Modules:

CE4051

Rationale and Purpose of the Module:

To give students an insight into deep learning and associated frameworks relevant to edge computing with sufficient practical work to enable them to implement their own deep learning systems. The module uses the Python programming language to showcase concepts and techniques such as batch normalization, pruning and quantization, the MobileNet architecture, hyperparameter tuning and optimization and other neural networks.

Syllabus:

1. Introduction to Deep Learning, Deep Neural Networks and Convolution Neural Networks. Training and inference. 2. Model optimization techniques such as network pruning and quantization. 8-bit quantization and Binary Neural Networks. 3. MobileNet V1/V2 architectures, Hyperparameter tuning. 4. Recurrent Neural Networks and applications. 5. Use of deep learning processor unit for edge applications. 6. Custom application development, including building the hardware, optimizing the trained model, and using the optimized model to accelerate a design.

Learning Outcomes:

Cognitive (Knowledge, Understanding, Application, Analysis, Evaluation, Synthesis)

On successful completion of this module, students will be able to: 1. Describe the key components in machine learning and deep learning systems. 2. Demonstrate best practices to build and train deep neural networks, identify architecture parameters, and apply deep learning solutions. 3. Design a Convolutional Neural Network, apply it to visual detection and recognition tasks. 4. Perform hyperparameter tuning and optimization for deep learning networks. 5. Build and train Recurrent Neural Networks (RNN).

Affective (Attitudes and Values)

On successful completion of this module, students will be able to: 1. Evaluate the limitations of current deep learning techniques. 2. Contribute meaningfully to team project development with deep learning systems.

Psychomotor (Physical Skills)

On successful completion of this module, students will be able to: N/A

How the Module will be Taught and what will be the Learning Experiences of the Students:

The module will be delivered using a blended learning approach using on-line lectures and labs. This module features a mix of theory, case studies and hands-on practical work to provide a high engagement learning experience of real-world deep learning techniques for edge computing applications.

Research Findings Incorporated in to the Syllabus (If Relevant):

Prime Texts:

Ian Goodfellow, Yoshua Bengio, Aaron Courville (2016) Deep Learning https://www.deeplearningbook.org/ , MIT Press

Stevens, Antiga, Viehmann (2020) Deep Learning with PyTorch , Manning Publications

Aurélien Géron (2020) Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, 2nd Edition , O'Reilly Media, Inc

Other Relevant Texts:

Programme(s) in which this Module is Offered:

Semester - Year to be First Offered:

Module Leader:

Brendan.Mullane@ul.ie