Recall that a density $\varphi$ is said to be $\epsilon$-biased if its correlation with every ${\mathbb F}_2$-linear function $f$ is at most $\epsilon$ in magnitude. In the lingo of pseudorandomness, one says that $\varphi$ fools the class of ${\mathbb F}_2$-linear functions:

[...]

## Recent comments

Ohad Klein: In 26, in the "Affine subspace partition" definition, "may b...Ohad Klein: In the very end of prop 12, I think there should be an index...Ohad Klein: In 25 (also in the book) "one one child".Noam Lifshitz: In corollary 17, should it be $\widehat{\mathrm{Maj}_n}(S) =...Ohad Klein: In 49 (56 in the book), it looks like a typo: $E[f_i(y^(j))]...Ryan O'Donnell: Hope so; I'm quite happy with it so far. (Thanks to all who...Yi Zhang: I got it now!!