3546 - Deep Learning

Delivery Options ON-LINE

Course Description

Extend your knowledge and understanding of Machine Learning to Deep Neural Networks. In this course we will cover the theory and practice of modern neural nets through a series of exercises and examples in different domains. You will build your own algorithms to classify images, perform rudimentary language translation and generate synthetic images or music.

Learning Outcomes

Know the theory and practice of modern neural networks.
Use Tensorflow2 to create and train deep neural networks
Tuning deep neural networks for different tasks
Understand the difference between various network architectures like CNN, RNN, transformer and generative algorithms
Apply deep-learning network architectures to solve a range of problems- e.g. classify images, predict trends and generate artworks

Prerequisites

3253 Machine Learning

Recommendations

You should have a laptop with at least 8 GB of RAM that can run recent Windows, Mac or Linux operating systems. You will need to have access to a laptop or desktop outside class with an i5 or preferably i7 processor, that can run recent Windows, Mac or Linux operating systems. Ideally the machine should also have an NVIDIA graphics card but this is not a requirement. Any software you’ll need is free and mostly open source. You will receive further instructions in class.