Deep Learning with Fast.ai
Piazza forum: https://piazza.com/class/k0av1h3l3ng3bp
Fast.ai aims to provide a very practical introduction to Deep Learning for people who already know how to code using Python, fast.ai library and Pytorch. This course covers a lot of application areas of Deep Learning as well as some more technical aspects of the neural network training process. Here is a random sample of topics covered in the course:
- Image classification with Deep Learning
- Image segmentation
- NLP and Tabular data
- Backpropagation and building Neural Networks from scratch
- Regularization and convolutions
- ResNet, U-Net, GANs
- Loss functions and optimizers
- Looking inside the model
- Advanced training techniques
We expect you to be comfortable with
- linear algebra,
- (matrix) calculus,
- probability theory,
- Python (on a good level)
This is going to be tough course, so do not take the prerequisites lightly.
We will use the reversed classroom approach, where we watch lectures at home and discuss them in the class. Discussions will be led by students themselves (we with Tambet, of course, will be present). Depending on the final number of attendees, students will be divided into pairs or bigger groups, which will be responsible for moderating discussion for a particular lecture. Also, these students will design either a short test or a simple homework based on the same lecture. Other students will complete the test or (and) submit the homework after the class. There will be a lot of self-organization in this course, be ready!
Semianrs are scheduled on Thursday 14.15 - 16.00 Liivi 2 - 512.
To pass the course students need to lead a discussion of one lecture, prepare the test/homework and score at least 60% of points from all other tests/homeworks.
Dmytro Fishman (dmytro at ut.ee) and Tambet Matiisen (tambet.matiisen at ut.ee) are the main instructors of the course Prof. Raul Vicente (raul.vicente.zafra at ut.ee) - coordinating the course