Introductory materials for AI
Last Updated by Gabriel Perdue on 2019/December/20
- The Google Machine Learning Crash Course:
- Deep Learning (book) - probably the best overall textbook in the field for a general practitioner:
- Textbook by Goodfellow, Bengio, and Courville
- http://www.deeplearningbook.org/
- For a complete PDF, visit: https://github.com/janishar/mit-deep-learning-book-pdf.git
- Practical Deep Learning for Coders:
- 7x 2-hour videos
- Theory is de-emphasized in exchange for a focus on power-use of the PyTorch and Fastai library
- https://course.fast.ai/
- Deep Learning from the Foundations:
- 7x 2-hour videos
- Recovers some of the theory skipped in “Practical Deep Learning for Coders”
- https://course.fast.ai/part2
- Other Fastai video courses, links, etc.
- https://distill.pub
- Distill collects essays on the underlying mechanics and math and pairs them with very nice visualizations of key concepts.
- Dive Into Deep Learning:
- http://d2l.ai/
- Interactive deep learning notebooks from a fundamental level based on NumPy.
- Stanford CS 231n
- http://cs231n.github.io
- This is a sophisticated course focusing specifically on Convolutional NNs - the software is out of date but the lecture material remains excellent.
- The TensorFlow web page is an excellent resource - most of their tutorial modules come with links to an executable Google Colab notebook, so they are easy to play with (with a GPU) on the web for free.
- Stanford CS 20
- https://web.stanford.edu/class/cs20si/syllabus.html
- This is a fairly compact course on how to use TensorFlow 1.X (code is not current with TF 2.X, so be careful of that) with a good GitHub repo of example code backing up the lecture slides.
- The PyTorch web page is also an excellent resource with tutorial notebooks executable in a Google Colab notebook for free