Analysis of Boolean Functions ( CMU 18 - 859 S , Spring 2007 ) Lecture 16 : The Hypercontractivity Theorem

نویسنده

  • Ryan O’Donnell
چکیده

1 Statement and history In this lecture, we prove the full-blown hypercontractivity theorem for {−1, 1} n. The idea behind the statement is that the T ρ operator smooths, or " reasonable-izes " , functions. Some examples, including two that were mentioned in the last lecture: Example 1.2 • q = 4, p = 2, ρ = 1/ √ 3: T 1/ √ 3 f 4 ≤ f 2 • q = 2, p = 4/3, ρ = 1/ √ 3: T 1/ √ 3 f 2 ≤ f 4/3 • q = q, p = 2, ρ = 1/ √ q − 1: T 1/ √ q−1 f 4 ≤ f 2 One corollary of the last of these is often quite sufficient; it's also a generalization of the original (2, 4, 1/ √ 3)-hypercontractivity result we proved easily by induction: Proof: f 2 q = d k=0 f =k 2 q = T 1/ √ q−1 d k=0 q − 1 k f =k 2 q ≤ d k=0 q − 1 k f =k 2 2 = d k=0 (q − 1) k |S|=kˆf (S) 2 ≤ (q − 1) d S ˆ f (S) 2 = (q − 1) d f 2 2 ,

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Analysis of Boolean Functions ( CMU 18 - 859 S , Spring 2007 ) Lecture 28 : Szemerédi ’ s Regularity

We investigated Schur triples (equivalently, triangles) when studying the BLR test for linearity in Lecture 2. There we considered picking x,y ∈ F2 at random, letting z = x + y, and checking whether f(x) + f(y) + f(z) = 0 (mod 2). If we were interested in triangle-freeness, we might instead check that f(x) = f(y) = f(z) = 1; equivalently, that f(x)f(y)f(z) = 1. Recall from Lecture 2 (and simila...

متن کامل

Analysis of Boolean Functions ( CMU 18 - 859 S , Spring 2007 ) Lecture 29 : Open Problems

Small total influence implies a large coefficient: Prove or disprove: For every f : {−1, 1} → {−1, 1} there exists some S such that |f̂(S)| ≥ 2−O(I(f)). One might also try to add the condition that the S satisfies |S| ≤ O(I(f)). A lower bound of 2−O(I(f)2) follows from Friedgut’s Theorem, and with this one can also get that |S| ≤ O(I(f)). The result definitely holds for monotone functions: From ...

متن کامل

Analysis of Boolean Functions ( CMU 18 - 859 S , Spring 2007 ) Lecture 14 : The KKL Theorem and Friedgut ’ s Theorem

As a special case, if a Boolean function is balanced, taking the value 1 on exactly half of its inputs, then f̂(∅) = Ex[f(x)] = 0, so Var(f) = 1 by the proposition. Hence Theorem 1.2 tells us that every balanced function has a variable with influence at least Ω(log n/n). In other words, not all variables can have small influence; if a function has, say, constant total influence, then there must ...

متن کامل

A Theorist ’ s Toolkit ( CMU 18 - 859 T , Fall 2013 ) Lecture 24 : Hardness Assumptions

This lecture is about hardness and computational problems that seem hard. Almost all of the theory of hardness is based on assumptions. We make assumptions about some problems, then we do reductions from one problem to another. Then, we want to make the minimal number of assumptions necessary to show computational hardness. In fact, all work on computational complexity and hardness is essential...

متن کامل

A Theorist ’ s Toolkit ( CMU 18 - 859 T , Fall 2013 ) Lecture 16 : Constraint Satisfaction Problems 10 / 30 / 2013

First let’s make sure this is actually a relaxation of our original problem. To see this, consider an optimal cut F ∗ : V → {1,−1}. Then, if we let ~ vi = (F (vi), 0, . . . , 0), all of the constraints are satisfied and the objective value remains the same. So, any solution to the original problem is also a solution to this vector program and therefore we have a relaxation. We will use SDPOpt t...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2007