§5.5: Highlight: Peres’s Theorem

Theorem 14 says that if $f$ is an unbiased linear threshold function $f(x) = \mathrm{sgn}(a_1 x_1 + \cdots + a_n x_n)$ in which all $a_i$’s are “small” then the noise stability $\mathbf{Stab}_\rho[f]$ is at least (roughly) $\frac{2}{\pi} \arcsin \rho$. Rephrasing in terms of noise sensitivity, this means $\mathbf{NS}_\delta[f]$ is at most (roughly) $\tfrac{2}{\pi} \sqrt{\delta} [...]

§5.4: Degree-1 weight

In this section we prove two theorems about the degree-$1$ Fourier weight of boolean functions: \[ \mathbf{W}^{1}[f] = \sum_{i=1}^n \widehat{f}(i)^2. \]


§5.3: The Fourier coefficients of Majority

In this section we will analyze the Fourier coefficients of $\mathrm{Maj}_n$. In fact, we give an explicit formula for them in Theorem 16 below. But most of the time this formula is not too useful; instead, it’s better to understand the Fourier coefficients of $\mathrm{Maj}_n$ asymptotically as $n \to \infty$.


§5.2: Majority, and the Central Limit Theorem

Majority is one of the more important functions in boolean analysis and its study motivates the introduction of one of the more important tools: the Central Limit Theorem (CLT).


§5.1: Linear threshold functions and polynomial threshold functions

Recall from Chapter 2.1 that a linear threshold function (abbreviated LTF) is a boolean-valued function $f : \{-1,1\}^n \to \{-1,1\}$ that can be represented as \begin{equation} \label{eqn:generic-LTF} f(x) = \mathrm{sgn}(a_0 + a_1 x_1 + \cdots + a_n x_n) \end{equation} for some constants $a_0, a_1, \dots, a_n \in {\mathbb R}$.


§4.5: Highlight: LMN’s work on constant-depth circuits

Having derived strong results about the Fourier spectrum of small DNFs and CNFs, we will now extend to the case of constant-depth circuits. We begin by describing how Håstad applied his Switching Lemma to constant-depth circuits. We then describe some Fourier-theoretic consequences coming from a very early (1989) work in analysis of boolean functions [...]

§4.4: Håstad’s Switching Lemma and the spectrum of DNFs

Let’s further investigate how random restrictions can simplify DNF formulas.


§4.3: Random restrictions

In this section we describe the method of applying random restrictions. This is a very “Fourier-friendly” way of simplifying a boolean function.


§4.2: Tribes

In this section we study the tribes DNF formulas, which serve as an important examples and counterexamples in analysis of boolean functions. Perhaps the most notable feature of the tribes function is that (for a suitable choice of parameters) it is essentially unbiased and yet all of its influences are quite tiny.


§4.1: DNF formulas

Perhaps the commonest way of representing a boolean function $f : \{0,1\}^n \to \{0,1\}$ is by a DNF formula: