Open course notes and teaching material that notes here actually draw on. This is a short shelf on purpose: a course earns a place by being cited, not by being famous.
Machine learning and computer vision
CS231n: Deep Learning for Computer Vision, Stanford University. The course notes are published openly and stand on their own as a text. They are the reference this garden reaches for whenever the question is mechanical rather than conceptual: how a loss is actually computed, how gradients actually flow backward, what a convolution does to a volume.
Cited by Train/Validation/Test (classification notes), Loss Functions (linear classification), Gradient Descent (optimization), Backpropagation (optimization part 2), Activation Functions and Artificial Neural Networks (neural networks part 1), CNNs and Pooling and CNN Architectures (convolutional networks).
Security and reliability
Fuzz Testing of Application Reliability, University of Wisconsin-Madison. The project page for the original fuzz work and the studies that followed it. Worth reading as primary material: fuzzing began as a class project testing whether UNIX utilities would survive random input, and many of them did not.
Cited by Fuzzing.
Engineering practice
Google Engineering Practices: Speed of Code Reviews. Google’s published guidance on review turnaround, and one of the few openly available documents that treats review latency as an engineering variable rather than a matter of manners.
Cited by Code Review.
Related Notes
Sources
- CS231n. https://cs231n.github.io/ . Title “CS231n Deep Learning for Computer Vision” and the statement that the notes accompany the Stanford CS class.
- University of Wisconsin-Madison. https://pages.cs.wisc.edu/~bart/fuzz/ . Title “Fuzz Testing of Application Reliability”.
- Google. https://google.github.io/eng-practices/review/reviewer/speed.html . Title “Speed of Code Reviews”.