+ All Categories
Home > Documents > Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the...

Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the...

Date post: 30-Mar-2015
Category:
Upload: marcel-lamkins
View: 217 times
Download: 0 times
Share this document with a friend
Popular Tags:
41
Invariance Principles in Theoretical Computer Science Ryan ’Donnell Carnegie Mellon University O
Transcript
Page 1: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Invariance Principles

in Theoretical Computer Science

Ryan ’Donnell

Carnegie Mellon University

O

Page 2: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

1. Describe some TCS results requiring

variants of the Central Limit Theorem.

Talk Outline

2. Show a flexible proof of the CLT

with error bounds.

3. Open problems and an advertisement.

Page 3: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

1. Describe some TCS results requiring

variants of the Central Limit Theorem.

Talk Outline

2. Show a flexible proof of the CLT

with error bounds.

3. Open problems and an advertisement.

Page 4: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Linear Threshold Functions

Page 5: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Linear Threshold Functions

Page 6: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Learning Theory [O-Servedio’08]

Thm: Can learn LTFs f in poly(n) time,

just from correlations E[f(x)xi].

Key:

when all |ci| ≤ ϵ.

Page 7: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Property Testing [Matulef-O-Rubinfeld-Servedio’09]

Thm: Can test if is

ϵ-close to an LTF with poly(1/ϵ) queries.

Key:

when all |ci| ≤ ϵ.

Page 8: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Derandomization [Meka-Zuckerman’10]

Thm: PRG for LTFs with seed

length O(log(n) log(1/ϵ)).

Key:

even when xi’s not fully independent.

Page 9: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Multidimensional CLT?

when all small compared to

For

Page 10: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Derandomization+ [Gopalan-O-Wu-Zuckerman’10]

Thm: PRG for “functions of O(1) LTFs”

with seed length O(log(n) log(1/ϵ)).

Key: Derandomized multidimensional CLT.

Page 11: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Property Testing+ [Blais-O’10]

Thm: Testing if is a

Majority of k bits needs kΩ(1) queries.

Key:

assuming E[Xi] = E[Yi], Var[Xi] = Var[Yi],

and some other conditions.

(actually, a multidimensional version)

Page 12: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Social Choice,Inapproximability [Mossel-O-Oleszkiewicz’05]

Thm: a) Among voting schemes where no

voter has unduly large influence,

Majority is most robust to noise.

b) Max-Cut is UG-hard to .878-approx.

Key: If P is a low-deg. multilin. polynomial,

assuming P has “small coeffs. on each coord.”

Page 13: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

1. Describe some TCS results requiring

variants of the Central Limit Theorem.

Talk Outline

2. Show a flexible proof of the CLT

with error bounds.

3. Open problems and an advertisement.

Page 14: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Gaussians

Standard Gaussian: G ~ N(0,1). Mean 0, Var 1.

a + bG also a “Gaussian”: N(a,b2)

Sum of independent Gaussians is Gaussian:

If G ~ N(a,b2), H ~ N(c,d2) are independent,

then G + H ~ N(a+c,b2+d2).

Anti-concentration: Pr[ G ∈ [u−ϵ, u+ϵ] ] ≤ O(ϵ).

Page 15: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

X1, X2, X3, … independent, ident. distrib.,

mean 0, variance σ2,

Central Limit Theorem (CLT)

Page 16: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

CLT with error bounds

X1 + · · · + Xnis “close to” N(0,1),

assuming Xi is not too wacky.

X1, X2, …, Xn independent, ident. distrib.,

mean 0, variance 1/n,

wacky:

Page 17: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Niceness of random variables

Say E[X] = 0, stddev[X] = σ.

eg: ±1. N(0,1). Unif on [-a,a].

not nice:

def: (≥ σ).

“def”: X is “nice” if

Page 18: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Niceness of random variables

Say E[X] = 0, stddev[X] = σ.

eg: ±1. N(0,1). Unif on [-a,a].

not nice:

def: (≥ σ).

def: X is “C-nice” if

Page 19: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Y “ϵ-close” to Z:

Berry-Esseen Theorem

X1, X2, …, Xn independent, ident. distrib.,

mean 0, variance 1/n,

X1 + · · · + Xnis ϵ-close to N(0,1),

assuming Xi is C-nice, where

[Shevtsova’07]: .7056

Page 20: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

General Case

X1, X2, …, Xn independent, ident. distrib.,

mean 0,

X1 + · · · + Xnis ϵ-close to N(0,1),

assuming Xi is C-nice,

Page 21: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Berry-Esseen: How to prove?

1. “Characteristic functions”

2. “Stein’s method”

3. “Replacement” = think like a cryptographer

X1, X2, …, Xn indep., mean 0,

S = X1 + · · · + XnG ~ N(0,1).ϵ-close to

Page 22: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Indistinguishability of random variables

S “ϵ-close” to G:

Page 23: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Indistinguishability of random variables

S “ϵ-close” to G:

u

Page 24: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Indistinguishability of random variables

S “ϵ-close” to G:

ut

Page 25: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Indistinguishability of random variables

S “ϵ-close” to G:

Page 26: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Replacement method

S “ϵ-close” to G:

Page 27: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Replacement method

X1, X2, …, Xn indep., mean 0,

S = X1 + · · · + Xn

G ~ N(0,1)

For smooth

Page 28: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Replacement method

X1, X2, …, Xn indep., mean 0,

G = G1 + · · · + Gn

For smooth

S = X1 + · · · + Xn

Hybrid argument

Page 29: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

X1, X2, …, Xn indep., mean 0,

SY = Y1 + · · · + Yn

For smooth

SX = X1 + · · · + Xn

Invariance principle

Y1, Y2, …, Yn Var[Xi] = Var[Yi] =

Page 30: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Hybrid argument

Def: Zi = Y1 + · · · + Yi + Xi+1 + · · · + Xn

SX = Z0, SY = Zn

X1, X2, …, Xn, Y1, Y2, …, Yn, independent,

matching means and variances.

SX = X1 + · · · + Xn SY = Y1 + · · · + Ynvs.

Page 31: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Hybrid argument

Zi = Y1 + · · · + Yi + Xi+1 + · · · + Xn

Goal:

X1, X2, …, Xn, Y1, Y2, …, Yn, independent,

matching means and variances.

Page 32: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Zi = Y1 + · · · + Yi + Xi+1 + · · · + Xn

Page 33: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Zi = Y1 + · · · + Yi + Xi+1 + · · · + Xn

where U = Y1 + · · · + Yi−1 + Xi+1 + · · · + Xn.

Note: U, Xi, Yi independent.

Goal:

Page 34: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

=

by indep. and matching means/variances!

Page 35: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Variant Berry-Esseen: Say

If X1, X2, …, Xn & Y1, Y2, …, Yn indep.

and have matching means/variances, then

Page 36: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Usual Berry-Esseen:

If X1, X2, …, Xn indep., mean 0,

Hack

Page 37: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Usual Berry-Esseen:

If X1, X2, …, Xn indep., mean 0,

Variant Berry-Esseen

+ Hack

Usual Berry-Esseen

except with error O(ϵ1/4)

Page 38: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Extensions are easy!

Vector-valued version:

Use multidimensional Taylor theorem.

Derandomized version:

If X1, …, Xm C-nice, 3-wise indep., then

X1+···+ Xm is O(C)-nice.

Higher-degree version:

X1, …, Xm C-nice, indep., Q is a deg.-d poly.,

then Q(X1, …, Xm) is O(C)d-nice.

Page 39: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

1. Describe some TCS results requiring

variants of the Central Limit Theorem.

Talk Outline

2. Show a flexible proof of the CLT

with error bounds.

3. Open problems, advertisement, anecdote?

Page 40: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Open problems

1. Recover usual Berry-Esseen via the

Replacement method.

2. Vector-valued: Get correct dependence

on test sets K. (Gaussian surface area?)

3. Higher-degree: improve (?) the

exponential dependence on degree d.

4. Find more applications in TCS.

Page 41: Ryan Donnell Carnegie Mellon University O. 1. Describe some TCS results requiring variants of the Central Limit Theorem. Talk Outline 2. Show a flexible.

Do you like LTFs and PTFs?

Do you like probability and geometry?

Oct. 21-22 (“just before FOCS”) workshopat the Princeton Intractability Center:

Analysis and Geometry of Boolean Threshold Functions

Diakonikolas! Kane! Meka! Rubinfeld! Servedio! Shpilka! Vempala! And more!

http://intractability.princeton.edu/blog/2010/08/workshop-ltfptf/


Recommended