Note I took this during summer semester; YMML. Very interesting subject that covers all the fundamentals of Deep Learning like basic feed forward networks, convolutional networks and recurrent networks. What interested me highly were the large amount of applications covered such as GANs, Variational Auto-Encoders, Adversarial ML, attention mechanisms and various SOTA architectures such as Transformers and BERT. The lecture slides themselves were easy to follow enough though the lecturer was quite difficult to understand. I found it better to just read the slides and get extra context online. The staff are helpful though, with tutors and lecturers communicating through forums frequently and giving helpful advice. The assignments were fun and engaging, having you implement various models along with special techniques using TensorFlow/keras. Exam was very fair; you won't encounter anything you haven't seen before, and you're given plenty of practise material in the form of weekly tutorial quizzes (not compulsory) and practise exams. Overall I'm satisfied with this unit; delivery of lectures could definitely be improved but content wise it's great.

Anonymous, Semester 2, 2021