Chapter 9 exercises

Continue reading Chapter 9 exercises

§9.6: Highlight: The Kahn–Kalai–Linial Theorem

Recalling the social choice setting of Chapter 2.5, consider a $2$-candidate, $n$-voter election using a monotone voting rule $f : \{-1,1\}^n \to \{-1,1\}$. We assume the impartial culture assumption (that the votes are independent and uniformly random), but with a twist: one of the candidates, say $b \in \{-1,1\}$, is able to secretly bribe $k$ voters, fixing their votes to $b$. (Since $f$ is monotone, this is always the optimal way for the candidate to fix the bribed votes.) How much can this influence the outcome of the election?
Continue reading §9.6: Highlight: The Kahn–Kalai–Linial Theorem

§9.5: Applications of hypercontractivity

With the $(2,q)$- and $(p,2)$-Hypercontractivity Theorems in hand, let’s revisit some applications we saw in Sections 1 and 2.
Continue reading §9.5: Applications of hypercontractivity

§9.4: Two-function hypercontractivity and induction

At this point we have established that if $f : \{-1,1\} \to {\mathbb R}$ then for any $p \leq 2 \leq q$, \[ \|\mathrm{T}_{\sqrt{p-1}} f\|_2 \leq \|f\|_p, \qquad \|\mathrm{T}_{1/\sqrt{q-1}} f\|_q \leq \|f\|_2. \] We would like to extend these facts to the case of general $f : \{-1,1\}^n \to {\mathbb R}$; i.e., establish the $(p,2)$- and $(2,q)$-Hypercontractivity Theorems stated at the beginning of the chapter. A natural approach is induction.
Continue reading §9.4: Two-function hypercontractivity and induction

§9.3: $(2,q)$- and $(p,2)$-hypercontractivity for a single bit

Although you can get a lot of mileage out of studying the $4$-norm of random variables, it’s also natural to consider other norms.
Continue reading §9.3: $(2,q)$- and $(p,2)$-hypercontractivity for a single bit

§9.2: Small subsets of the hypercube are noise-sensitive

An immediate consequence of the Bonami Lemma is that for any $f : \{-1,1\}^n \to {\mathbb R}$ and $k \in {\mathbb N}$, \begin{equation} \label{eqn:2-4-hypercon-k} \|\mathrm{T}_{1/\sqrt{3}} f^{=k}\|_4 = \tfrac{1}{\sqrt{3}^k} \|f^{=k}\|_4 \leq \|f^{=k}\|_2. \end{equation}
Continue reading §9.2: Small subsets of the hypercube are noise-sensitive

§9.1: Low-degree polynomials are reasonable

As anyone who has worked in probability knows, a random variable can sometimes behave in rather “unreasonable” ways. It may be never close to its expectation. It might exceed its expectation almost always, or almost never. It might have finite $1$st, $2$nd, and $3$rd moments, but an infinite $4$th moment. All of this poor behaviour can cause a lot of trouble — wouldn’t it be nice to have a class of “reasonable” random variables?
Continue reading §9.1: Low-degree polynomials are reasonable

Chapter 9: Basics of hypercontractivity

In 1970, Bonami proved the following central result:

The Hypercontractivity Theorem Let $f : \{-1,1\}^n \to {\mathbb R}$ and let ${1 \leq p \leq q \leq \infty}$. Then $\|\mathrm{T}_\rho f\|_q \leq \|f\|_p$ for $0 \leq \rho \leq \sqrt{\tfrac{p-1}{q-1}}$.

Continue reading Chapter 9: Basics of hypercontractivity

Chapter 8 notes

The origins of the orthogonal decomposition described in Section 3 date back to the work of Hoeffding [Hoe48] (see also [vMis47]).
Continue reading Chapter 8 notes

Chapter 8 exercises, continued

Continue reading Chapter 8 exercises, continued