Introduction to Deep Learning (Lecture with Project)
| Lecturer | Hyemin Ahn |
| Allocation to curriculum | See TUMonline |
| Offered in | Wintersemester 2021/22 |
| Semester weekly hours | 4 |
| Scheduled dates | See TUMonline |
| Contact | Hyemin Ahn (hyemin.ahn@tum.de) |
Content
This course will cover the following topics in terms of (1) theoretical background, and (2) practical implemtation based on python3 and pytorch.
1. Artificial Neural Network (ANN), Optimization, Backpropagation.
2. Overfitting and Performance Validation
3. Convolutional Neural Network, AlexNet, VGG, and ResNet
4. Natural Language Processing, Transformer
5. Generative Models, VAE, GAN.
6. Neural Style Transfer.
7. Project Proposal and Presentation
Previous Knowledge Expected
Fundamentals of Linear Algebra, Probability and Statistics, Optimization.
Basic python will be dealt in course briefly, but it is recommended to have programming skills in Python3.
Objective
At the end of this course, students are able to:
- To build a background knowledge for reading and understanding deep learning based conference/journal papers related to one's own research interest.
- To design and train a deep neural network which is appropriate to solve one's own research problem based on the PyTorch.