A new central limit theorem and decomposition for Gaussian polynomials, with an application to deterministic approximate counting

Anindya De; Rocco A. Servedio

doi:10.1007/s00440-017-0804-y

Probability Theory and Related Fields

August 2018, Volume 171, Issue 3–4, pp 981–1044 | Cite as

A new central limit theorem and decomposition for Gaussian polynomials, with an application to deterministic approximate counting

Authors
Authors and affiliations

Anindya De
Rocco A. Servedio

Article

First Online: 30 September 2017

Received: 19 March 2016

Revised: 25 April 2017

216 Downloads

Abstract

One of the main results of this paper is a new multidimensional central limit theorem (CLT) for multivariate polynomials under Gaussian inputs. Roughly speaking, the new CLT shows that any collection of Gaussian polynomials with small eigenvalues (suitably defined) must have a joint distribution which is close to a multidimensional Gaussian distribution. The new CLT is proved using tools from Malliavin calculus and Stein’s method. A second main result of the paper, which complements the new CLT, is a new decomposition theorem for low-degree multilinear polynomials over Gaussian inputs. Roughly speaking, this result shows that (up to some small error) any such polynomial is very close to a polynomial which can be decomposed into a bounded number of multilinear polynomials all of which have extremely small eigenvalues. An important feature of this decomposition theorem is the delicate control obtained between the number of polynomials in the decomposition versus their eigenvalues. As the main application of these results, we give a deterministic algorithm for approximately counting satisfying assignments of a degree-d polynomial threshold function (PTF) over the domain $\{-1,1\}^n$; this is a well-studied problem from theoretical computer science. More precisely, given as input a degree-d polynomial $p(x_1,\dots ,x_n)$ over ${{\mathbb {R}}}^n$ and a parameter $\epsilon > 0$, the algorithm approximates

$$\begin{aligned} \mathop {\mathbf{Pr}}_{x \sim \{-1,1\}^n}[p(x) \ge 0] \end{aligned}$$

to within an additive $\pm \epsilon $ in time $O_{d,\epsilon }(1)\cdot \mathrm {poly}(n^d)$. (Since it is NP-hard to determine whether the above probability is nonzero, any sort of efficient multiplicative approximation is almost certainly impossible even for randomized algorithms.) Note that the running time of the algorithm (as a function of $n^d$, the number of coefficients of a degree-d PTF) is a fixed polynomial. The fastest previous algorithm for this problem (Kane, CoRR, arXiv:1210.1280, 2012), based on constructions of unconditional pseudorandom generators for degree-d PTFs, runs in time $n^{O_{d,c}(1) \cdot \epsilon ^{-c}}$ for all $c > 0$.

Keywords

Gaussian chaos Central limit theorem Polynomials Malliavin calculus Stein’s method Regularity lemma

Mathematics Subject Classification

60F05 68Q87

This is a preview of subscription content, to check access.

Notes

Acknowledgements

We thank Ilias Diakonikolas for his contributions in the early stages of this project. We also thank Rafal Latala, Michel Ledoux, Elchanan Mossel, Ivan Nourdin and Krzysztof Oleszkiewicz for answering questions about the CLT. Part of this work was done when A.D. was hosted by Oded Regev and the Simons Institute. A.D. would like to thank them for their kind hospitality and support.

Appendix: A dealing with non-multilinear polynomials

The decomposition procedure that we use relies heavily on the fact that the input polynomials $p_i$ are multilinear. To handle general (non-multilinear) degree-d polynomials, the first step of our algorithm is to transform them to (essentially) equivalent multilinear degree-d polynomials. This is accomplished by a simple procedure whose performance is described in the following theorem.⁴ Note that given Theorem 55, we assume that the polynomial p given as input in Theorem 3 is multilinear.

Theorem 55

There is a deterministic procedure Multilinearize with the following properties: The algorithm takes as input a (not necessarily multilinear) variance-1 degree-d polynomial p over ${{\mathbb {R}}}^n$ and an accuracy parameter $\delta > 0$. It runs in time $O_{d,\delta }(1)\cdot poly(n^d)$ and outputs a multilinear degree-d polynomial q over $R^{n'}$, with $n' \le O_{d,\delta }(1) \cdot n$, such that

$$\begin{aligned} \left| \mathbf{Pr}_{x \sim N(0,1)^n}[p(x) \ge 0] - \mathbf{Pr}_{x \sim N(0,1)^{n'}}[q(x) \ge 0] \right| \le O(\delta ). \end{aligned}$$

Proof

The procedure Multilinearize is given below. (Recall that a diagonal entry of a q-tensor $f = \sum _{i_1,\dots ,i_q=1} f(i_1, \dots , i_q)\cdot e_{i_1} \otimes \cdots \otimes e_{i_q}$ is a coefficient $f(i_1,\dots ,i_q)$ that has $i_a=i_b$ for some $a \ne b.$)

It is clear that all the tensors $g_j$ are multilinear, so by Remark 11 the polynomial q that the procedure Linearize outputs is multilinear. The main step in proving Theorem 55 is to bound the variance of $\tilde{q}-q$, so we will establish the following claim:

Claim 56

$\mathbf{Var}[\tilde{q}-q] \le {\frac{d}{K}} \cdot \mathbf{Var}[\tilde{q}].$

Proof

We first observe that

$$\begin{aligned} \mathbf{Var}[\tilde{q}-q] = \mathbf{Var}\left[ \sum _{j=1}^d (I_j(\tilde{f}_j) - I_j(g_j)) \right] = \sum _{j=1}^d \mathbf{E}\left[ (I_j(\tilde{f}_j) - I_j(g_j))^2\right] , \end{aligned}$$

(54)

where the first equality is because $g_0=f_0$ and the second is by Claim 12 and the fact that each $I_j(\tilde{f}_j), I_j(g_j)$ has mean 0 for $j \ge 1.$ Now fix a value $1 \le j \le d.$ Since each $g_j$ is obtained from $\tilde{f}_j$ by zeroing out diagonal elements, again using Claim 12 we see that

$$\begin{aligned} \mathbf{E}\left[ (I_j(\tilde{f}_j) - I_j(g_j))^2\right] = j! \cdot \Vert \tilde{f}_j - g_j\Vert ^2_F, \end{aligned}$$

(55)

where the squared Frobenius norm $\Vert \tilde{f}_j - g_j\Vert ^2_F$ equals the sum of squared entries of the tensor $\tilde{f}_j - g_j$. Now,observe that the entry $\alpha _{i_1,\dots ,i_j}=f_j(i_1, \ldots , i_j)$ of the tensor $f_j $ maps to the entry

$$\begin{aligned}&\alpha _{i_1, \ldots , i_j} \frac{(e_{i_1,1} + \cdots + e_{i_1, K})}{\sqrt{K}} \otimes \ldots \otimes \frac{(e_{i_j,1} + \cdots +e_{i_j, K})}{\sqrt{K}}\\&\quad = \alpha _{i_1, \ldots , i_j} \cdot \sum _{(\ell _1, \ldots , \ell _{{j}}) \in [K]^{{j}}} \frac{1}{K^{{j}/2}} \otimes _{{{a=1}}}^{{j}} e_{i_{{a}},\ell _{{a}}} \end{aligned}$$

when $\tilde{q}$ is constructed from p. Further observe that all $K^j$ outcomes of $\otimes _{{{a=1}}}^{{j}} e_{i_{{a}},\ell _{{a}}}$ are distinct. Since $g_j$ is obtained by zeroing out the diagonal entries of $\tilde{f}_j$, we get that

$$\begin{aligned} \Vert \tilde{f}_j - g_j \Vert _F^2 = \sum _{{(i_1, \ldots , i_j) \in [n]^j}} (\alpha _{i_1, \ldots , i_j})^2 \cdot \frac{1}{K^j} \cdot |{\mathcal {S}}_{K, j}| \end{aligned}$$

where the set ${\mathcal {S}}_{K,j} = \{{(\ell _1, \ldots , \ell _j) \in [K]^j : \ell _1, \ldots , \ell _j} \text { are not all distinct}\}$. It is easy to see that $|{\mathcal {S}}_{K, j}| \le (j^2 \cdot K^j)/{K}$, so we get

$$\begin{aligned} \Vert \tilde{f}_j - g_j \Vert _F^2 \le \sum _{{(i_1, \ldots , i_j) \in [n]^j}} (\alpha _{i_1, \ldots , i_j})^2 \cdot \frac{j^2}{{K}}. \end{aligned}$$

Returning to (54) and (55), this yields

$$\begin{aligned} \mathbf{Var}[q-\tilde{q}] \le \sum _{j=1}^d j! \cdot \sum _{{(i_1, \ldots , i_j) \in [n]^j}} (\alpha _{i_1, \ldots , i_j})^2 \cdot \frac{j^2}{{K}} \le \frac{d^2}{K} \cdot \left( \sum _{j=1}^d j! \cdot \sum _{{(i_1, \ldots , i_j) \in [n]^j}} (\alpha _{i_1, \ldots , i_j})^2\right) . \end{aligned}$$

Using Fact 13 and Claim 12, we see that

$$\begin{aligned} \mathbf{Var}[p] = \sum _{j=1}^d \mathbf{Var}[I_j(f_j)] = \sum _{j=1}^d \mathbf{E}[I_j(f_j)^2] = \sum _{j=1}^d j! \cdot \sum _{{(i_1, \ldots , i_j) \in [n]^j}} (\alpha _{i_1, \ldots , i_j})^2. \end{aligned}$$

It is easy to see that $\mathbf{Var}[\tilde{q}]=\mathbf{Var}[p]$, which equals 1 by assumption, so we have that $\mathbf{Var}[q-\tilde{q}]\le {\frac{d^2}{K}} \cdot \mathbf{Var}[\tilde{q}]$ as desired. $\square $

To finish the proof of Theorem 55, observe that by our choice of K we have $\mathbf{Var}[q-\tilde{q}] \le (\delta /d)^{3d} \cdot \mathbf{Var}[ \tilde{q}]$. Since $q-\tilde{q}$ has mean 0 and $\mathbf{Var}[\tilde{q}]=1$ we may apply Lemma 6, and we get that $|\mathbf{Pr}_{x \sim N(0,1)^{n'}}[q(x) \ge 0] - \mathbf{Pr}_{x \sim N(0,1)^{n'}}[\tilde{q}(x) \ge 0]| \le O(\delta ).$ The theorem follows by observing that the two distributions $p(x)_{x \sim N(0,1)^n}$ and $\tilde{q}(x)_{x \sim N(0,1)^{n'}}$ are identical. $\square $

References

1.
Aaronson, S.: P $\stackrel{?}{=}$ NP. http://www.scottaaronson.com/papers/pnp.pdf. Earlier version in “Open Problems in Mathematics”, Nash, J.F., Jr., Rassias, M.Th. (eds.) (2017)
2.
Alon, N., Babai, L., Itai, A.: A fast and simple randomized algorithm for the maximal independent set problem. J. Algorithms 7, 567–583 (1985)MathSciNetCrossRefzbMATHGoogle Scholar
3.
Aziz, H., Paterson, M., Leech, D.: Efficient algorithm for designing weighted voting games. In: IEEE International Multitopic Conference, pp. 1–6 (2007)Google Scholar
4.
Ajtai, M., Wigderson, A.: Deterministic simulation of probabilistic constant depth circuits. In: Proceedings of 26th IEEE Symposium on Foundations of Computer Science (FOCS), pp. 11–19 (1985)Google Scholar
5.
Adamczak, R., Wolff, P.: Concentration inequalities for non-Lipschitz functions with bounded derivatives of higher order. Probab. Theory Relat. Fields 162(3), 531–586 (2015)MathSciNetCrossRefzbMATHGoogle Scholar
6.
Bhatia, R.: Matrix Analysis. Springer, Basel (2000)zbMATHGoogle Scholar
7.
Breuer, P., Major, P.: Central limit theorems for non-linear functionals of Gaussian fields. J. Multivar. Anal. 13(3), 425–441 (1983)CrossRefzbMATHGoogle Scholar
8.
Chatterjee, S.: A new method of normal approximation. Ann. Probab. 36(4), 1584–1610, 07 (2008)MathSciNetCrossRefzbMATHGoogle Scholar
9.
Chatterjee, S.: Fluctuations of eigenvalues and second-order Poincaré inequalities. Probab. Theory Relat. Fields 143, 1–40 (2009)CrossRefzbMATHGoogle Scholar
10.
Cartwright, D., Sturmfels, B.: The number of eigenvalues of a tensor. Linear Algebra Appl. 432(2), 942–952 (2013)MathSciNetCrossRefzbMATHGoogle Scholar
11.
Carbery, A., Wright, J.: Distributional and $L^q$ norm inequalities for polynomials over convex bodies in $R^n$. Math. Res. Lett. 8(3), 233–248 (2001)MathSciNetCrossRefzbMATHGoogle Scholar
12.
De, A., Diakonikolas, I., Feldman, V., Servedio, R.: Near-optimal solutions for the Chow parameters problem and low-weight approximation of halfspaces. In: Proceedings of 44th ACM Symposium on Theory of Computing (STOC), pp. 729–746 (2012)Google Scholar
13.
De, A., Diakonikolas, I., Servedio, R.A.: The inverse Shapley value problem. ICALP 1, 266–277 (2012)MathSciNetzbMATHGoogle Scholar
14.
De, A., Diakonikolas, I., Servedio, R.: Deterministic approximate counting for degree-2 polynomial threshold functions. Manuscript (2013)Google Scholar
15.
De, A., Diakonikolas, I., Servedio, R.: Deterministic approximate counting for juntas of degree-2 polynomial threshold functions. Manuscript (2013)Google Scholar
16.
Diakonikolas, I., Harsha, P., Klivans, A., Meka, R., Raghavendra, P., Servedio, R.A., Tan, L.-Y.: Bounding the average sensitivity and noise sensitivity of polynomial threshold functions. In: STOC, pp. 533–542 (2010)Google Scholar
17.
Diakonikolas, I., Kane, D.M., Nelson, J.: Bounded independence fools degree-2 threshold functions. In: Proceedings of 51st IEEE Symposium on Foundations of Computer Science (FOCS), pp. 11–20 (2010)Google Scholar
18.
Dowson, D.C., Landau, B.V.: The Frechet distance between multivariate normal distributions. J. Multivar. Anal. 12(3), 450–455 (1982)MathSciNetCrossRefzbMATHGoogle Scholar
19.
Diakonikolas, I., O’Donnell, R., Servedio, R., Wu, Y.: Hardness results for agnostically learning low-degree polynomial threshold functions. In: SODA, pp. 1590–1606 (2011)Google Scholar
20.
Diakonikolas, I., Servedio, R., Tan, L.-Y., Wan, A.: A regularity lemma, and low-weight approximators, for low-degree polynomial threshold functions. In: CCC, pp. 211–222 (2010)Google Scholar
21.
Feller, W.: An Introduction to Probability Theory and Its Applications. Wiley, London (1968)zbMATHGoogle Scholar
22.
Friedman, J., Wigderson, A.: On the second eigenvalue of hypergraphs. Combinatorica 15(1), 43–65 (1995)MathSciNetCrossRefzbMATHGoogle Scholar
23.
Goldmann, M., Håstad, J., Razborov, A.: Majority gates vs. general weighted threshold gates. Comput. Complex. 2, 277–300 (1992)MathSciNetCrossRefzbMATHGoogle Scholar
24.
Gopalan, P., Klivans, A., Meka, R., Stefankovic, D., Vempala, S., Vigoda, E.: An fptas for #knapsack and related counting problems. In: FOCS, pp. 817–826 (2011)Google Scholar
25.
Golub, G., Van Loan, C.F.: Matrix Computations. The Johns Hopkins University Press, Baltimore (1996)zbMATHGoogle Scholar
26.
Gopalan, P., Meka, R., Reingold, O.: DNF sparsification and a faster deterministic counting algorithm. Comput. Complex. 22(2), 275–310 (2013)MathSciNetCrossRefzbMATHGoogle Scholar
27.
Gopalan, P., O’Donnell, R., Wu, Y., Zuckerman, D.: Fooling functions of halfspaces under product distributions. In: IEEE Conference on Computational Complexity (CCC), pp. 223–234 (2010)Google Scholar
28.
Håstad, J.: On the size of weights for threshold gates. SIAM J. Discret. Math. 7(3), 484–492 (1994)MathSciNetCrossRefGoogle Scholar
29.
Janson, S.: Gaussian Hilbert Spaces. Cambridge University Press, Cambridge (1997)CrossRefzbMATHGoogle Scholar
30.
Kane, D.M.: The Gaussian surface area and noise sensitivity of degree-d polynomial threshold functions. In: CCC, pp. 205–210 (2010)Google Scholar
31.
Kane, D.M.: k-independent gaussians fool polynomial threshold functions. In: IEEE Conference on Computational Complexity, pp. 252–261 (2011)Google Scholar
32.
Kane, D.M.: A small PRG for polynomial threshold functions of gaussians. In: FOCS, pp. 257–266 (2011)Google Scholar
33.
Kane, D.M.: The correct exponent for the Gotsman–Linial conjecture. arXiv:1210.1283 (2012)
34.
Kane, D.M.: A pseudorandom generator for polynomial threshold functions of gaussian with subpolynomial seed length. arXiv:1210.1280 (2012)
35.
Kane, D.M.: A structure theorem for poorly anticoncentrated gaussian chaoses and applications to the study of polynomial threshold functions. In: FOCS, pp. 91–100 (2012)Google Scholar
36.
Kalai, A., Klivans, A., Mansour, Y., Servedio, R.: Agnostically learning halfspaces. SIAM J. Comput. 37(6), 1777–1805 (2008)MathSciNetCrossRefzbMATHGoogle Scholar
37.
Karnin, Z.S., Rabani, Y., Shpilka, A.: Explicit dimension reduction and its applications. SIAM J. Comput. 41(1), 219–249 (2012)MathSciNetCrossRefzbMATHGoogle Scholar
38.
Latala, R.: Estimates of moments and tails of gaussian chaoses. Ann. Probab. 34(6), 2315–2331 (2006)MathSciNetCrossRefzbMATHGoogle Scholar
39.
Latala, R.: Personal communication (2013)Google Scholar
40.
Ledoux, M.: Personal communication (2013)Google Scholar
41.
Luby, M., Velickovic, B.: On deterministic approximation of DNF. Algorithmica 16(4/5), 415–433 (1996)MathSciNetCrossRefzbMATHGoogle Scholar
42.
Luby, M., Velickovic, B., Wigderson, A.: Deterministic approximate counting of depth-2 circuits. In: Proceedings of the 2nd ISTCS, pp. 18–24 (1993)Google Scholar
43.
Myhill, J., Kautz, W.: On the size of weights required for linear-input switching functions. IRE Trans. Electron. Comput. EC10(2), 288–290 (1961)CrossRefGoogle Scholar
44.
Mossel, E., O’Donnell, R., Oleszkiewicz, K.K.: Noise stability of functions with low influences: invariance and optimality. Ann. Math. 171, 295–341 (2010)MathSciNetCrossRefzbMATHGoogle Scholar
45.
Minsky, M., Papert, S.: Perceptrons: An Introduction to Computational Geometry. MIT Press, Cambridge (1968)zbMATHGoogle Scholar
46.
Muroga, S., Toda, I., Takasu, S.: Theory of majority switching elements. J. Frankl. Inst. 271, 376–418 (1961)CrossRefzbMATHGoogle Scholar
47.
Muroga, S.: Threshold Logic and Its Applications. Wiley-Interscience, New York (1971)zbMATHGoogle Scholar
48.
Meka, R., Zuckerman, D.: Pseudorandom generators for polynomial threshold functions. http://arxiv.org/abs/0910.4122 (2009)
49.
Meka, R., Zuckerman, D.: Pseudorandom generators for polynomial threshold functions. In: STOC, pp. 427–436 (2010)Google Scholar
50.
Naor, J., Naor, M.: Small-bias probability spaces: efficient constructions and applications. SIAM J. Comput. 22(4), 838–856 (1993). (Earlier version in STOC’90) MathSciNetCrossRefzbMATHGoogle Scholar
51.
Nourdin, I.: Lectures on gaussian approximations with Malliavin calculus. Technical report. http://arxiv.org/abs/1203.4147v3, 28 June 2012
52.
Nourdin, I.: Personal communication (2013)Google Scholar
53.
Nualart, D., Peccati, G.: Central limit theorems for sequences of multiple stochastic integrals. Ann. Probab. 33(1), 177–193 (2005)MathSciNetCrossRefzbMATHGoogle Scholar
54.
Nourdin, I., Peccati, G.: Stein’s method meets Malliavin calculus: a short survey with new estimates. Technical report. http://arxiv.org/abs/0906.4419v2, 17 Sep 2009
55.
Nourdin, I., Peccati, G., Réveillac, A.: Multivariate normal approximation using Stein’s method and Malliavin calculus. Ann. Inst. H. Poincaré Probab. Stat. 46(1), 45–58 (2010)MathSciNetCrossRefzbMATHGoogle Scholar
56.
Oleszkiewicz, K.: Personal communication (2013)Google Scholar
57.
Orponen, P.: Neural networks and complexity theory. In: Proceedings of the 17th International Symposium on Mathematical Foundations of Computer Science, pp. 50–61 (1992)Google Scholar
58.
Podolskii, V.V.: Perceptrons of large weight. Probl. Inf. Transm. 45(1), 46–53 (2009)MathSciNetCrossRefzbMATHGoogle Scholar
59.
Servedio, R.: Every linear threshold function has a low-weight approximator. Comput. Complex. 16(2), 180–209 (2007)MathSciNetCrossRefzbMATHGoogle Scholar
60.
Sherstov, A.A.: Halfspace matrices. Comput. Complex. 17(2), 149–178 (2008)MathSciNetCrossRefzbMATHGoogle Scholar
61.
Sherstov, A.: The intersection of two halfspaces has high threshold degree. In: Proceedings of 50th IEEE Symposium on Foundations of Computer Science (FOCS) (2009)Google Scholar
62.
Shalev-Shwartz, S., Shamir, O., Sridharan, K.: Learning kernel-based halfspaces with the 0–1 loss. SIAM J. Comput. 40(6), 1623–1646 (2011)MathSciNetCrossRefzbMATHGoogle Scholar
63.
Trevisan, L.: A note on approximate counting for $k$-DNF. In: Proceedings of the Eighth International Workshop on Randomization and Computation, pp. 417–426 (2004)Google Scholar
64.
Viola, E.: The sum of $d$ small-bias generators fools polynomials of degree $d$. Comput. Complex. 18(2), 209–217 (2009)MathSciNetCrossRefzbMATHGoogle Scholar

Copyright information

Authors and Affiliations

Anindya De
- 1
Email authorView author's OrcID profile
Rocco A. Servedio
- 2

1.Northwestern UniversityEvanstonUSA
2.Columbia UniversityNew YorkUSA

A new central limit theorem and decomposition for Gaussian polynomials, with an application to deterministic approximate counting

Abstract

Keywords

Mathematics Subject Classification

Notes

Acknowledgements

References

Copyright information

Authors and Affiliations

Personalised recommendations

Cite article