This is an undergraduate-level introductory course for deep learning. There have been enormous advances in the field of artificial intelligence over the past few decades, especially based on deep learning. However, it is not easy to see what frontiers the current AI is facing and what underlying methods are used to enable these advances. This course aims to provide an overview of traditional/emerging topics and applications in deep learning, and basic skill sets to understand/implement some of the latest algorithms. Some topics that will be covered in this course include:
- Visual recognition (image classification, object detection, semantic segmentation).
- Natural language understanding (machine translation, image captioning, visual question answering).
- Generative models (image/audio/text synthesis).
- Self-supervised and unsupervised learning
- Planning and RL (searching algorithms, MDP)
To understand each topic, we will learn some basics in deep learning, such as various neural network architectures, loss functions, optimization techniques, etc., together with machine learning libraries to implement these ideas, such as Pytorch.
This is an introductory class for general topics in AI and deep learning, and we encourage the students who want to learn more on each specific topic to take the advanced courses (e.g. CS484, CS494, CS576).
Prof. Seunghoon Hong: [email protected]
E3-1, Room 3429
Jaehoon Yoo: [email protected]
Jinwoo Kim: [email protected]
Ian Goodfellow, Yoshua Bengio, Aron Courville, Deep Learning, MIT Press (Electronic copy)
Undergraduate-level courses for linear algebra, discrete mathematics, machine learning (optional), statistics (optional)
Students are expected to be familiar with Python programming (we encourage learning basics of Python if not).
Attendance: 5%
Quiz: 15%
Assignment: 40%
Final project: 40%