This is arguably the best comprehensive resource available. Written by Marc Peter Deisenroth, A. Aldo Faisal, and Cheng Soon Ong, this book bridges the gap between high school math and advanced ML concepts.
This is the single most important concept in ML. The gradient is a vector containing all the partial derivatives. It points in the direction of the steepest ascent.
Pitfall 1: Confusing derivative with gradient. calculus for machine learning pdf link
Pitfall 2: Forgetting the constant multiple rule.
Pitfall 3: Chain Rule confusion in Backprop. This is arguably the best comprehensive resource available
print(numerical_derivative(lambda x: x**2, 3)) # Output: ~6.0
There are many textbooks, but most are too academic and dense for programmers. Below are the best free resources available online (legally free via open-source licenses). Pitfall 2: Forgetting the constant multiple rule
In Machine Learning, the derivative tells you: If I change this weight slightly, how much does the error change?