Download presentation
Presentation is loading. Please wait.
1
Putting a Junta to the Test Joint work with Eldar Fischer, Dana Ron, Shmuel Safra, and Alex Samorodnitsky Guy Kindler
2
Property Testing o P – property o f – input o Goal: Distinguish, using the fewest possible queries, between f has P f is -far from having P d(f,g) = Pr x [ f(x)≠g(x) ]
3
History o Testing Proofs (PCP): BLR o Combinatorial properties: GGR o PRS: Logic AND, monotonous DNF.
4
Juntas Boolean Functions: f(f( )= n entries
5
Juntas Boolean Functions: 11111 1 11 f(f( )=
6
Juntas 11111 1 11 f(f( )= j-junta: depends on at most j coordinates. 11111 1 11
7
Juntas 1111 1 1 1 1 f(f( )= 111 11 j-junta: depends on at most j coordinates.
8
1 1111 111 11 f(f( )= Definition of j-Junta Test 11111 1 11 111 1 111 11111 11 1 1 1 1 11 1 1 111 1 1 1 11 11 1 1 f(f( )= f(f( f(f( f(f( 1 1 1
9
1 1111 111 11 f(f( )= 11111 1 11 111 1 111 11111 11 1 1 1 1 11 1 1 111 1 1 1 11 11 1 1 f(f( )= f(f( f(f( f(f( 1 1 1 Accept? Reject? Definition of j-Junta Test
10
Before we test juntas … Given a set I of coordinates, can we verify that f does not depend on it?
11
I-independence test
12
I 1 1 111 1 1 1 111 1
13
I-independence test I w f(f( )= f(f( 1 1 111 1 1 1 111 1 1 1 1 1 1 1
14
I-independence test I f(f( )= w f(f( 1 1 111 1 1 1 111 1 1 1 1 1 1 1 Claim: If Pr[ I is detected] ≤ then f is at most ” -dependent on I ” g independent of I, d(f,g)≤ variation f (I)
15
Claim: If Pr[ I is detected] ≤ then f is at most ” -dependent on I ” I-independence test I f(f( )= w f(f( 1 1 111 1 1 1 111 1 1 1 1 1 1 1 g independent of I, d(f,g)≤ variation f (I)
16
Claim: If Pr[ I is detected] ≤ then f is at most ” -dependent on I ” I-independence test Proof: Let g(w z 0 )=Maj z {f(w z)} Define p(w)=Pr z [f(w z) ≠g(w z)] variation f (I) I f(f( )= w f(f( 1 1 111 1 1 1 111 1 1 1 1 1 1 1 g independent of I, d(f,g)≤ variation f (I)
17
1. Partition the coordinates into r subsets. The j-Junta Test I1I1 IrIr r=10j 2 A j-junta is independent of all but j subsets !
18
1. Partition the coordinates into r subsets. 2. Run the independence-test r/ times on each subset. The j-Junta Test I1I1 IrIr If f has variation “ /r” on a subset, it is almost surely detected!
19
1. Partition the coordinates into r subsets. 2. Run the independence-test r/ times on each subset. 3. Accept if ≤j of the subsets are detected. The j-Junta Test I1I1 IrIr Completeness: Soundness: If f is far from being a junta, then the test rejects with probability ½. Soundness: If f is far from being a junta, then the test rejects with probability ½.
20
Lemma: For every Boolean f, unless f is –close to a j-junta, w.h.p., the test rejects. Soundness at least j+1 subsets have variation /r J [n], | J |≤j, variation f ([n]\ J )< over the partitions of [n],
21
Variations 1
22
I
23
I
24
We’ll prove that unless f is –close to a j-junta, w.h.p. the test rejects. at least j+1 subsets have variation /r | J |≤j, and variation f ([n]\ J )< over the partitions of [n], For t /r, let I1I1 IrIr If | J |>j, easy !!
25
We’ll prove that unless f is –close to a j-junta, w.h.p. the test rejects. at least j+1 subsets have variation /r | J |≤j, and variation f ([n]\ J )< over the partitions of [n], Fix t /r and let I1I1 IrIr Assume variation f ([n]\ J )> . Then !!! Claim: w.h.p.
26
o Recall: For each i in Claim: w.h.p.
27
I J o Recall: For each i in Claim: w.h.p.
28
I J The Unique-Variation
29
I J
30
I J
31
I J
32
I J
33
I J
34
I J
35
I J Q.E.D
36
Other Results o Shown number of queries: j 4 / o Using adaptivity: j 3 / o Using two-sidedness: j 2 / o Allowing (2j)-juntas: j 2 / o Variables in General probability spaces. o “f” is “g” test, where g is a j-junta. o Lower Bound: at least (j) 1/2 queries are needed
37
Open Problems o Improve lower bound to j 2 / (perhaps via random-walk convergence on Z 2 ) o “f is g” for non-juntas? o Characterize efficiently testable properties via Fourier transform??
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.