Some typos / errors found by CSC 580 Fall 2022 students at University of Arizona

[zcc1307]

Hi Hal,

Thanks a lot for making this wonderful textbook publicly available!

Here is a collection of possible errors students found (with my verification) when I used this book to teach a machine learning course at the University of Arizona. I haven't checked all existing bug reports; hopefully some of the issues we found are helpful for improving the book! (Special thanks to Jesse Chen, Thang Duong, Oghenevovwe Abiodun Ikumariegbe, Hao Qin, Winston Zeng for reporting these).

Best,
Chicheng

Chapter 3:

In Fig 3.4 and the relevant paragraphs, the positives/negatives mentioned should be reversed

In Fig 3.16 and 3.17 and relevant paragraphs, blue sphere => red sphere, grey spheres => green spheres.

A minor mistake on Page 31, paragraph 4, "Consider Figure 3.3.". There's no Figure 3.3.
The author was referring to Figure 3.4 instead.

page 32, paragraph 2, "If you consider the 3-nearest neighbors of the test point in Figure 3.4, you will see that two of them are positive and one is negative. Through voting, positive would win."
Based on the figure, it should be “two of them are negative and one is positive. Through voting, negative would win.”

Chapter 4:

pg44:
(The -> This) is actually a bad idea.
pg46:
This makes complete sense geometrically, since all that matters is which side of the plane a test point falls on, (now -> not) how far it is from that plane.
This distance along w is exactly the (activiation -> activation) of that example, with no bias.
pg47:
Fig 4.8 is empty
In the rest of this book, we’ll refer to the weight vector, and to (the) hyperplane it defines, interchangeably
pg49:
If I give you a data set and the hyperplane that separates (itthen -> it, then) the margin is the distance between the hyperplane and the nearest point.

Chapter 5:

Several figures are missing

Chapter 7:

pg94: gradient ascent -> gradient descent
pg98: “Av is well defined if A is DxM” -> “Av is well defined if A is MxD”; the subsequent matrix-vector product definition needs to be adjusted as well

Chapter 11:

pg141: Eq. (11.1) -- the last entry of the fourth row should be x_3 x_D
pg143: “algorithm 29” should be “algorithm 11.2”

Chapter 15:
pg179: “It must be the case that | x_n - \mu_b | \leq | x_n - \mu_b |” -> “It must be the case that | x_n - \mu_b | \leq | x_n - \mu_a |”

Chapter 16:

page 190 and 191: (16.16) is missing a square in the numerator and a factor of D in the denominator. Same holds for (16.22).