PHS 597: Deep learning with applications to genomics and health informatics.

Deep learning is called the “new electricity” for modern science and technology. Getting a good understanding of modern deep learning methods would be critical for statisticians who are interested in “big data” research. 2023 was probably one of the most notable years in recent history where AI becomes a daily buzzword. ChatGPT and similar tools afterwards completely change how people view AI. Not a single day goes by without us hearing something about AI. For us as data scientists and biostatisticians, we need to know AI to be successful in our future careers.

This course is jointly taught by Drs. Sen Yang and Dajiang Liu. It covers basic theory and applications for modern deep learning methods in genomics and health informatics. We will survey basic theory for deep neural networks, convoluted neural networks, sequence models, transformer, generative adversarial networks, graphical neural networks, diffusion models, etc. Given the rapid development of the field, this year’s course will undergo major changes compared to the materials from earlier years.

We will largely follow the book “Understanding Deep Learning” [https://udlbook.github.io/udlbook/]. It is nicely written, contains modern materials, and has exercises. Other useful sources include:

1. Deep Learning book by Ian Goodfellow and Yoshua Bengio and Aaron Courville (link): This is one of the first deep learning books, which we followed in earlier years. The material is getting obsolete and the book is not well organized. But it remains a useful reference, particularly for its earlier part on math theories and optimizations.
2. Probablistic machine learning [https://probml.github.io/pml-book/book1.html].

Notes.

  1. Shallow/deep neural networks (slides; notes: 01/22)
  2. Computation for neural networks (slides)
  3. Regularization (slides) (notes: 02/10 ).
  4. Convoluted neural networks (slides) (notes: 02/19 02/26 03/03)
  5. Sequence models (slides) (notes: 03/26 ).
  6. Attention and transformer (slides) (notes: 03/31)
  7. Generative models (slides)
  8. Graph neural network (slides)
  9. Reinforcement learning (slides)