Kuatnya pemisahan junta


16

Kita mengatakan bahwa fungsi Boolean adalah -junta jika memiliki paling banyak mempengaruhi variabel.f : { 0 , 1 } n{ 0 , 1 } f:{0,1}n{0,1}k kf fkk

Biarkan menjadi -junta. Nyatakan variabel dengan . Perbaiki Jelas, ada sedemikian rupa sehingga mengandung setidaknya dari variabel yang mempengaruhi .f : { 0 , 1 } n{ 0 , 1 } f:{0,1}n{0,1}2 k 2kf fx 1 , x 2 , ... , x n x1,x2,,xnS 1 = { x 1 , x 2 , ... , x n2 },S 2 = { x n2+1,xn2+2,,xn}.

S1={x1,x2,,xn2},S2={xn2+1,xn2+2,,xn}.
S{S1,S2}S{S1,S2}SSkkff

Sekarang mari , dan anggap bahwa adalah -far dari setiap -junta (yaitu, kita harus mengubah sebagian kecil dari setidaknya dari nilai-nilai untuk membuatnya menjadi -junta). Bisakah kita membuat versi "kuat" dari pernyataan di atas? Yaitu, adakah konstanta universal , dan himpunan sedemikian rupa sehingga adalah -far dari setiap fungsi yang paling banyak berisi variabel yang mempengaruhi di ?ϵ>0ϵ>0f:{0,1}n{0,1}f:{0,1}n{0,1}ϵϵ2k2kϵϵff2k2kccS{S1,S2}S{S1,S2}ffϵcϵckkSS

Catatan: Dalam rumusan asli pertanyaan, ditetapkan sebagai . Contoh Neal menunjukkan bahwa nilai itu tidak cukup. Namun, karena dalam pengujian properti kami biasanya tidak terlalu peduli dengan konstanta, saya sedikit mengendurkan kondisinya.cc22cc


Bisakah Anda menjelaskan persyaratan Anda? Apakah variabel "memengaruhi" kecuali nilai f selalu independen terhadap variabel? Apakah "mengubah nilai " berarti, mengubah salah satu nilai untuk beberapa tertentu ? fff(x)f(x)xx
Neal Young

Tentu saja, variabel x ixi memengaruhi jika ada string n-n bit yy sehingga f ( y ) f ( y )f(y)f(y) , di mana y y adalah string yy dengan koordinat ke - i -nya iterbalik. Mengubah nilai ff berarti membuat perubahan dalam tabel kebenarannya.

Jawaban:


17

Jawabannya iya". Buktinya dengan kontradiksi.

Untuk kenyamanan notasi, marilah kita menunjukkan variabel n / 2 pertama n/2dengan xx dan variabel n / 2 kedua n/2oleh yy . Misalkan f ( x , y )f(x,y) adalah δ-δ tutup ke fungsi f 1 ( x , y )f1(x,y) yang hanya bergantung pada koordinat kk dari xx . Nyatakan koordinatnya yang berpengaruh oleh T 1T1 . Demikian pula, anggaplah f ( x , y )f(x,y) adalahδ-δ tutup ke fungsi f 2 ( x , y )f2(x,y) yang hanya bergantung padakoordinat kk dari yy . Nyatakan koordinatnya yang berpengaruh oleh T 2T2 . Kita perlu membuktikan bahwa ff adalah 4 δ4δ - mendekati 2 k-2k junta ˜ f ( x , y )f~(x,y) .

Katakanlah ( x 1 , y 1 ) ( x 2 , y 2 ) jika x 1 dan x 2 menyetujui semua koordinat di T 1 dan y 1 dan y 2 setuju pada semua koordinat di T 2 . Kami memilih perwakilan acak dari setiap kelas kesetaraan secara acak. Biarkan ( ˉ x , ˉ y ) menjadi perwakilan untuk kelas ( x ,(x1,y1)(x2,y2)x1x2T1y1y2T2(x¯,y¯)y ) . Tentukan ˜ f sebagai berikut: ˜ f ( x , y ) = f ( ˉ x , ˉ y ) .(x,y)f~

f~(x,y)=f(x¯,y¯).

Jelaslah bahwa ˜ f adalah 2 k- junta (hanya bergantung pada variabel dalam T 1T 2 ) . Kami akan membuktikan bahwa itu pada jarak 4 δ dari f dalam harapan.f~2kT1T2)4δf

Kami ingin membuktikan bahwa Pr ˜ f ( Pr x , y ( ˜ f ( x , y ) f ( x , y ) ) )) = Pr ( f ( ˉ x , ˉ y ) f ( x , y ) ) 4 δ , di mana x dan y dipilih secara seragam secara acak. Pertimbangkan vektor acak

Prf~(Prx,y(f~(x,y)f(x,y)))=Pr(f(x¯,y¯)f(x,y))4δ,
xy~ X diperoleh darixdengan menjaga semua bit diT1dan secara acak membalik semua bit tidak diT1, dan vektor ~ y didefinisikan sama. Perhatikan bahwa Pr( ˜ f (x,y)f(x,y))=Pr(f( ˉ x , ˉ y )f(x,y))=Prx~xT1T1y~( f ( ˜ x , ˜ y ) f ( x , y ) ) .
Pr(f~(x,y)f(x,y))=Pr(f(x¯,y¯)f(x,y))=Pr(f(x~,y~)f(x,y)).

Kami memiliki, Pr ( f ( x , y ) f ( ˜ x , y ) ) Pr ( f ( x , y ) f 1 ( x , y ) ) + Pr ( f 1 ( x , y ) f 1 ( ˜ x , y ) ) + Pr ( f1 ( ˜ x , y ) f ( ˜ x , y ) ) δ + 0 + δ = 2 δ .

Pr(f(x,y)f(x~,y))Pr(f(x,y)f1(x,y))+Pr(f1(x,y)f1(x~,y))+Pr(f1(x~,y)f(x~,y))δ+0+δ=2δ.

Demikian pula, Pr ( f ( ˜ x , y ) f ( ˜ x , ˜ y ) ) 2 δ . Kami memiliki Pr ( f ( ˉ x , ˉ y ) f ( x , y ) ) 4 δ . QEDPr(f(x~,y)f(x~,y~))2δ

Pr(f(x¯,y¯)f(x,y))4δ.

Mudah untuk “membatalkan pengacakan” bukti ini. Untuk setiap ( x , y ) , misalkan ˜ f ( x , y ) = 1 jika f ( x , y ) = 1 untuk sebagian besar ( x , y ) dalam kelas ekivalensi ( x , y ) , dan ˜ f ( x , y ) = 0 , jika tidak.(x,y)f~(x,y)=1f(x,y)=1(x,y)(x,y)f~(x,y)=0


12

The smallest cc that the bound holds for is c=1212.41c=1212.41.

Lemmas 1 and 2 show that the bound holds for this cc. Lemma 3 shows that this bound is tight.

(In comparison, Juri's elegant probabilistic argument gives c=4c=4.)

Let c=121c=121. Lemma 1 gives the upper bound for k=0k=0.

Lemma 1: If ff is ϵgϵg-near a function gg that has no influencing variables in S2S2, and ff is ϵhϵh-near a function hh that has no influencing variables in S1S1, then ff is ϵϵ-near a constant function, where ϵ(ϵg+ϵh)/2cϵ(ϵg+ϵh)/2c.

Proof. Let ϵϵ be the distance from ff to a constant function. Suppose for contradiction that ϵϵ does not satisfy the claimed inequality. Let y=(x1,x2,,xn/2)y=(x1,x2,,xn/2) and z=(xn/2+1,,xn)z=(xn/2+1,,xn) and write ff, gg, and hh as f(y,z)f(y,z), g(y,z)g(y,z) and h(y,z)h(y,z), so g(y,z)g(y,z) is independent of zz and h(y,z)h(y,z) is independent of yy.

(I find it helpful to visualize ff as the edge-labeling of the complete bipartite graph with vertex sets {y}{y} and {z}{z}, where gg gives a vertex-labeling of {y}{y}, and hh gives a vertex-labeling of {z}{z}.)

Let g0g0 be the fraction of pairs (y,z)(y,z) such that g(y,z)=0g(y,z)=0. Let g1=1g0g1=1g0 be the fraction of pairs such that g(y,z)=1g(y,z)=1. Likewise let h0h0 be the fraction of pairs such that h(y,z)=0h(y,z)=0, and let h1h1 be the fraction of pairs such that h(y,z)=1h(y,z)=1.

Without loss of generality, assume that, for any pair such that g(y,z)=h(y,z)g(y,z)=h(y,z), it also holds that f(y,z)=g(y,z)=h(y,z)f(y,z)=g(y,z)=h(y,z). (Otherwise, toggling the value of f(y,z)f(y,z) allows us to decrease both ϵgϵg and ϵhϵh by 1/2n1/2n, while decreasing the ϵϵ by at most 1/2n1/2n, so the resulting function is still a counter-example.) Say any such pair is ``in agreement''.

The distance from ff to gg plus the distance from ff to hh is the fraction of (x,y)(x,y) pairs that are not in agreement. That is, ϵg+ϵh=g0h1+g1h0ϵg+ϵh=g0h1+g1h0.

The distance from ff to the all-zero function is at most 1g0h01g0h0.

The distance from ff to the all-ones function is at most 1g1h11g1h1.

Further, the distance from ff to the nearest constant function is at most 1/21/2.

Thus, the ratio ϵ/(ϵg+ϵh)ϵ/(ϵg+ϵh) is at most min(1/2,1g0h0,1g1h1)g0h1+g1h0,

min(1/2,1g0h0,1g1h1)g0h1+g1h0,
where g0,h0[0,1]g0,h0[0,1] and g1=1g0g1=1g0 and h1=1h0h1=1h0.

By calculation, this ratio is at most 12(21)=c/212(21)=c/2. QED

Lemma 2 extends Lemma 1 to general kk by arguing pointwise, over every possible setting of the 2k2k influencing variables. Recall that c=121c=121.

Lemma 2: Fix any kk. If ff is ϵgϵg-near a function gg that has kk influencing variables in S2S2, and ff is ϵhϵh-near a function hh that has kk influencing variables in S1S1, then ff is ϵϵ-near a function ˆff^ that has at most 2k2k influencing variables, where ϵ(ϵg+ϵh)/2cϵ(ϵg+ϵh)/2c.

Proof. Express ff as f(a,y,b,z)f(a,y,b,z) where (a,y)(a,y) contains the variables in S1S1 with aa containing those that influence hh, while (b,z)(b,z) contains the variables in S2S2 with bb containing those influencing gg. So g(a,y,b,z)g(a,y,b,z) is independent of zz, and h(a,y,b,z)h(a,y,b,z) is independent of yy.

For each fixed value of aa and bb, define Fab(y,z)=f(a,y,b,z)Fab(y,z)=f(a,y,b,z), and define GabGab and HabHab similarly from gg and hh respectively. Let ϵgabϵgab be the distance from FabFab to GabGab (restricted to (y,z)(y,z) pairs). Likewise let ϵhabϵhab be the distance from FabFab to HabHab.

By Lemma 1, there exists a constant cabcab such that the distance (call it ϵabϵab) from FabFab to the constant function cabcab is at most (ϵhab+ϵgab)/(2c)(ϵhab+ϵgab)/(2c). Define ˆf(a,y,b,z)=cabf^(a,y,b,z)=cab.

Clearly ˆff^ depends only on aa and bb (and thus at most kk variables).

Let ϵˆfϵf^ be the average, over the (a,b)(a,b) pairs, of the ϵabϵab's, so that the distance from ff to ˆff^ is ϵˆfϵf^.

Likewise, the distances from ff to gg and from ff to hh (that is, ϵgϵg and ϵh)ϵh) are the averages, over the (a,b)(a,b) pairs, of, respectively, ϵgabϵgab and ϵhabϵhab.

Since ϵab(ϵhab+ϵgab)/(2c)ϵab(ϵhab+ϵgab)/(2c) for all a,ba,b, it follows that ϵˆf(ϵg+ϵh)/(2c)ϵf^(ϵg+ϵh)/(2c). QED

Lemma 3 shows that the constant cc above is the best you can hope for (even for k=0k=0 and ϵ=0.5ϵ=0.5).

Lemma 3: There exists ff such that ff is (0.5/c)(0.5/c)-near two functions gg and hh, where gg has no influencing variables in S2S2 and hh has no influencing variables in S1S1, and ff is 0.50.5-far from every constant function.

Proof. Let yy and zz be xx restricted to, respectively, S1S1 and S2S2. That is, y=(x1,,xn/2)y=(x1,,xn/2) and z=(xn/2+1,,xn)z=(xn/2+1,,xn).

Identify each possible yy with a unique element of [N][N], where N=2n/2N=2n/2. Likewise, identify each possible zz with a unique element of [N][N]. Thus, we think of ff as a function from [N]×[N][N]×[N] to {0,1}{0,1}.

Define f(y,z)f(y,z) to be 1 iff max(y,z)12Nmax(y,z)12N.

By calculation, the fraction of ff's values that are zero is (12)2=12(12)2=12, so both constant functions have distance 1212 to f.

Define g(y,z) to be 1 iff y12N. Then g has no influencing variables in S2. The distance from f to g is the fraction of pairs (y,z) such that y<12N and z12N. By calculation, this is at most 12(112)=0.5/c

Similarly, the distance from f to h, where h(y,z)=1 iff z12N, is at most 0.5/c.

QED


First of all, thanks Neal! This indeed sums it up for k=0, and sheds some light on the general problem. However in the case of k=0 the problem is a bit degenerate (as 2k=k), so I'm more curious regarding the case of k1. I didn't manage to extend this claim for k>0, so if you have an idea on how to do it - I'd appreciate it. If it simplifies the problem, then the exact constants are not crucial; that is, ϵ/2-far can be replaced by ϵ/c-far, for some universal constant c.

2
I've edited it to add the extension to general k. And Yuri's argument below gives a slightly looser factor with an elegant probabilistic argument.
Neal Young

Sincere thanks Neal! This line of reasoning is quite enlightening.
Dengan menggunakan situs kami, Anda mengakui telah membaca dan memahami Kebijakan Cookie dan Kebijakan Privasi kami.
Licensed under cc by-sa 3.0 with attribution required.