Random Arithmetic Formulas Can Be Reconstructed Efficiently

Informally stated, we present here a randomized algorithm that given blackbox access to the polynomial f computed by an unknown/hidden arithmetic formula φ reconstructs, on average, an equivalent or smaller formula φ in time polynomial in the size of its output φ. Specifically, we consider arithmetic formulas wherein the underlying tree is a complete binary tree, the leaf nodes are labelled by affine forms (i.e. degree one polynomials) over the input variables and where the internal nodes consist of alternating layers of addition and multiplication gates. We call these alternating normal form (ANF) formulas. If a polynomial f can be computed by an arithmetic formula μ of size s, it can also be computed by an ANF formula φ, possibly of slightly larger size sO(1). Our algorithm gets as input blackbox access to the output polynomial f (i.e. for any point x in the domain, it can query the blackbox and obtain f(x) in one step) of a random ANF formula φ of size s (wherein the coefficients of the affine forms in the leaf nodes of φ are chosen independently and uniformly at random from a large enough subset of the underlying field). With high probability (over the choice of coefficients in the leaf nodes), the algorithm efficiently (i.e. in time sO(1)) computes an ANF formula φ of size s computing f. This then is the strongest model of arithmetic computation for which a reconstruction algorithm is presently known, albeit efficient in a distributional sense rather than in the worst case.

[1]  Gerhard Pfister,et al.  Primary Decomposition: Algorithms and Comparisons , 1997, Algorithmic Algebra and Number Theory.

[2]  David A. Cox,et al.  Ideals, Varieties, and Algorithms , 1997 .

[3]  Amir Yehudayoff,et al.  Arithmetic Circuits: A survey of recent results and open questions , 2010, Found. Trends Theor. Comput. Sci..

[4]  Erich Kaltofen,et al.  Computing with polynomials given by straight-line programs II sparse factorization , 1985, 26th Annual Symposium on Foundations of Computer Science (sfcs 1985).

[5]  Satyanarayana V. Lokam,et al.  Efficient Reconstruction of Random Multilinear Formulas , 2011, 2011 IEEE 52nd Annual Symposium on Foundations of Computer Science.

[6]  M. Ben-Or,et al.  A Deterministic Algorithm for Sparse Multivariate Polynominal Interpolation (Extended Abstract) , 1988, Symposium on the Theory of Computing.

[7]  Daniel Lazard,et al.  Thirty years of Polynomial System Solving, and now? , 2009, J. Symb. Comput..

[8]  Nader H. Bshouty,et al.  Size-depth tradeoffs for algebraic formulae , 1991, [1991] Proceedings 32nd Annual Symposium of Foundations of Computer Science.

[9]  Stasys Jukna,et al.  Boolean Function Complexity Advances and Frontiers , 2012, Bull. EATCS.

[10]  Amir Shpilka.  Interpolation of Depth-3 Arithmetic Circuits with Two Multiplication Gates , 2009, SIAM J. Comput..

[11]  J. Kollár.  Sharp effective Nullstellensatz , 1988 .

[12]  Satyanarayana V. Lokam,et al.  Reconstruction of depth-4 multilinear circuits with top fan-in 2 , 2012, STOC '12.

[13]  Eyal Kushilevitz,et al.  Learning functions represented as multiplicity automata , 2000, JACM.

[14]  Vikraman Arvind,et al.  New Results on Noncommutative and Commutative Polynomial Identity Testing , 2008, 2008 23rd Annual IEEE Conference on Computational Complexity.

[15]  Neeraj Kayal,et al.  Approaching the Chasm at Depth Four , 2013, Computational Complexity Conference.

[16]  Neeraj Kayal,et al.  Affine projections of polynomials , 2011, Electron. Colloquium Comput. Complex..

[17]  Joe W. Harris,et al.  Algebraic Geometry: A First Course , 1995 .

[18]  Erich Kaltofen,et al.  Computing with Polynomials Given By Black Boxes for Their Evaluations: Greatest Common Divisors, Factorization, Separation of Numerators and Denominators , 1990, J. Symb. Comput..

[19]  Erich Kaltofen,et al.  Improved Sparse Multivariate Polynomial Interpolation Algorithms , 1988, ISSAC.

[20]  Michael E. Saks,et al.  Minimizing Disjunctive Normal Form Formulas and AC0 Circuits Given a Truth Table , 2008, SIAM J. Comput..

[21]  Erich Kaltofen,et al.  Factorization of Polynomials Given by Straight-Line Programs , 1989, Adv. Comput. Res..

[22]  Linda Sellie,et al.  Exact learning of random DNF over the uniform distribution , 2009, STOC '09.

[23]  Teo Mora,et al.  Local Decomposition Algorithms , 1990, AAECC.

[24]  Yishay Mansour,et al.  Learning Boolean Functions via the Fourier Transform , 1994 .

[25]  Thomas W. Dubé.  A Combinatorial Proof of the Effective Nullstellensatz , 1993, J. Symb. Comput..

[26]  Nader H. Bshouty,et al.  Interpolating Arithmetic Read-Once Formulas in Parallel , 1998, SIAM J. Comput..

[27]  Juan Sabia,et al.  Effective equidimensional decomposition of affine varieties , 2002 .

[28]  Daniel A. Spielman,et al.  Randomness efficient identity testing of multivariate polynomials , 2001, STOC '01.

[29]  Gian-Carlo Rota,et al.  Apolarity and Canonical Forms for Homogeneous Polynomials , 1993, Eur. J. Comb..

[30]  Chee-Keng Yap,et al.  Fundamental problems of algorithmic algebra , 1999 .

[31]  Adam R. Klivans,et al.  Learning Arithmetic Circuits via Partial Derivatives , 2003, COLT.

[32]  Daniel Lazard,et al.  Solving systems of algebraic equations , 2001, SIGS.

[33]  Johan Håstad.  Tensor Rank is NP-Complete , 1990, J. Algorithms.

[34]  Erich Kaltofen,et al.  Polynomial-Time Reductions from Multivariate to Bi- and Univariate Integral Polynomial Factorization , 1985, SIAM J. Comput..

[35]  Heinz Kredel,et al.  Gröbner Bases: A Computational Approach to Commutative Algebra , 1993 .

[36]  Neeraj Kayal,et al.  Efficient algorithms for some special cases of the polynomial equivalence problem , 2011, SODA '11.

[37]  David Buchfuhrer,et al.  The Complexity of Boolean Formula Minimization , 2008, ICALP.

[38]  Matthias Aschenbrenner.  Ideal membership in polynomial rings over the integers , 2003, math/0305172.

[39]  Lance Fortnow,et al.  Efficient Learning Algorithms Yield Circuit Lower Bounds , 2006, COLT.

[40]  Andrew Wan,et al.  Mansour's Conjecture is True for Random DNF Formulas , 2010, COLT.

[41]  G. Greuel,et al.  A Singular Introduction to Commutative Algebra , 2002 .

[42]  K. Kalorkoti,et al.  A Lower Bound for the Formula Size of Rational Functions , 1982, SIAM J. Comput..

[43]  Neeraj Kayal,et al.  An exponential lower bound for homogeneous depth four arithmetic circuits with bounded bottom fanin , 2012, Electron. Colloquium Comput. Complex..

[44]  Jin-Yi Cai,et al.  Circuit minimization problem , 2000, STOC '00.

[45]  Neeraj Kayal.  Affine projections of polynomials: extended abstract , 2012, STOC '12.

[46]  Amir Shpilka,et al.  Reconstruction of Generalized Depth-3 Arithmetic Circuits with Bounded Top Fan-in , 2009, 2009 24th Annual IEEE Conference on Computational Complexity.

[47]  Richard Zippel,et al.  Interpolating Polynomials from Their Values , 1990, J. Symb. Comput..

[48]  David A. Cox,et al.  Using Algebraic Geometry , 1998 .

[49]  Grete Hermann,et al.  The question of finitely many steps in polynomial ideal theory , 1998, SIGS.

[50]  Grete Hermann,et al.  Die Frage der endlich vielen Schritte in der Theorie der Polynomideale , 1926 .

[51]  Teresa Krick,et al.  Membership problem, Representation problem and the Computation of the Radical for one-dimensional Ideals , 1991 .

[52]  T. Mignon,et al.  A quadratic bound for the determinant and permanent problem , 2004 .

[53]  Daniel Lazard,et al.  Resolution des Systemes d'Equations Algebriques , 1981, Theor. Comput. Sci..

[54]  Joachim von zur Gathen.  Permanent and determinant , 1987 .