COSIW06: Resources

Learn about Python

All of the major deep learning frameworks rely on Python, so you should become at least a little familiar with the basic use. There are many tutorials:

• An example
• Many more examples linked here.

Also you probably should become a little familiar with Jupyter notebooks, since that is the easy way to manage your deep learning code/processes. Again there are many ways to learn it:

• An example Jupyter tutorial.
• Google Colaboratory (aka "Colab") hosts free GPU-backed computation in a Jupyter-like version of a Google Doc, which is a great way to get started doing python. You can link Colab to your own Google Drive (to get data in and out). This article has tips on how to use Google Colaboratory's free GPU-backed deep learning for a variety of kinds of problems.

Learn about deep learning

• Before the seminar beins, please watch this video of Glassner's crash course on deep learning, recorded at SIGGRAPH 2018.
  • Running Time: 3 hours. Note: there is a gap in the middle when they take a break; it resumes at 1:48.
  • Description: Deep learning is a revolutionary technique for discovering patterns from data. We'll see how this technology works and what it offers us for computer graphics. There won't be any math. Attendees will learn how to use these tools to power their own creative and practical investigations and applications.
• Glassner also has a two-volume book on deep learning available on Kindle here and here – an equally intuitive approach to the topic as in that video course above.
• Nielson offers this free online book, which I found to be very helpful both for understanding the many mathematical concepts behind deep learning, and also with practical advice on how to navigate the very broad space of possible deep learning architectures.
• The lectures from Stanford CS231n: Convolutional Neural Networks for Visual Recognition are excellent and are all available online. They focus on visual recongnition, but the concepts apply in our area too.

Learn about audio processing

We will start the semester with a brief introduction to audio, so you won't really need to get a head-start on this. Nevertheless, we will gather some materials here.

• This course looks great but will start in mid September, which may be too late for us.
• Aatish Bhatia's blog post on the Fourier transform.
• Visualizing sound waves here and here.
• Live microphone oscilloscope
• Fourier approximation demo
• Harmonics demo
• Spectrogram visualizer

Free audio processing software

• Spear - FFT resynthesis application
• Audacity - audio editing application