Derandomizing Dynamic Programming and Beyond

We consider probabilistic circuits working over the real numbers, and using arbitrary semialgebraic functions of bounded description complexity as gates. We show that such circuits can be simulated by deterministic circuits with an only polynomial blowup in size. An algorithmic consequence is that randomization cannot substantially speed up dynamic programming. In arithmetic circuits, randomization cannot spare even one single gate.

[1]  Norbert Sauer,et al.  On the Density of Families of Sets , 1972, J. Comb. Theory A.

[2]  M. Jerrum,et al.  Bounding the Vapnik-Chervonenkis Dimension of Concept Classes Parameterized by Real Numbers , 2004, Machine Learning.

[3]  Joos Heintz,et al.  Corrigendum: Definability and Fast Quantifier Elimination in Algebraically Closed Fields , 1983, Theor. Comput. Sci..

[4]  Stasys Jukna.  Lower Bounds for Tropical Circuits and Dynamic Programs , 2014, Theory of Computing Systems.

[5]  Friedhelm Meyer auf der Heide,et al.  Simulating Probabilistic by Deterministic Algebraic Computation Trees , 1985, Theor. Comput. Sci..

[6]  Stasys Jukna,et al.  Tropical Complexity, Sidon Sets, and Dynamic Programming , 2016, SIAM J. Discret. Math..

[7]  David Haussler,et al.  Decision Theoretic Generalizations of the PAC Model for Neural Net and Other Learning Applications , 1992, Inf. Comput..

[8]  A. Tarski.  A Decision Method for Elementary Algebra and Geometry , 2023 .

[9]  Joos Heintz,et al.  Description of the connected components of a semialgebraic set in single exponential time , 1994, Discret. Comput. Geom..

[10]  Devdatt P. Dubhashi,et al.  Concentration of Measure for the Analysis of Randomized Algorithms: Contents , 2009 .

[11]  Michael Ben-Or,et al.  A theorem on probabilistic constant depth Computations , 1984, STOC '84.

[12]  Nicolai Vorobjov,et al.  Counting connected components of a semialgebraic set in subexponential time , 1992, computational complexity.

[13]  George E. Collins,et al.  Quantifier elimination for real closed fields by cylindrical algebraic decomposition , 1975 .

[14]  H. Warren.  Lower bounds for approximation by nonlinear manifolds , 1968 .

[15]  Dima Grigoriev.  Complexity lower bounds for randomized computation trees over zero characteristic fields , 1999, computational complexity.

[16]  Paul W. Goldberg,et al.  PAC-learning geometrical figures , 1992 .

[17]  Leonard M. Adleman,et al.  Two theorems on random polynomial time , 1978, 19th Annual Symposium on Foundations of Computer Science (sfcs 1978).

[18]  E. Szemerédi,et al.  Sorting inc logn parallel steps , 1983 .

[19]  Stasys Jukna,et al.  Greedy can also beat pure dynamic programming , 2018, Electron. Colloquium Comput. Complex..

[20]  Oded Goldreich,et al.  In a World of P=BPP , 2010, Studies in Complexity and Cryptography.

[21]  Volker Weispfenning,et al.  The Complexity of Linear Problems in Fields , 1988, Journal of symbolic computation.

[22]  Jacob T. Schwartz,et al.  Fast Probabilistic Algorithms for Verification of Polynomial Identities , 1980, J. ACM.

[23]  Udi Manber,et al.  The complexity of problems on probabilistic, nondeterministic, and alternating decision trees , 1985, JACM.

[24]  J. Milnor.  On the Betti numbers of real varieties , 1964 .

[25]  P. Assouad.  Densité et dimension , 1983 .

[26]  Dima Grigoriev,et al.  Complexity of Deciding Tarski Algebra , 1988, J. Symb. Comput..

[27]  John Gill,et al.  Relative to a Random Oracle A, PA != NPA != co-NPA with Probability 1 , 1981, SIAM J. Comput..

[28]  Oscar H. Ibarra,et al.  Fast Approximation Algorithms for the Knapsack and Sum of Subset Problems , 1975, JACM.

[29]  Marc Snir,et al.  Lower Bounds on Probabilistic Linear Decision Trees , 1985, Theor. Comput. Sci..

[30]  Murali K. Ganapathy,et al.  On the number of zero-patterns of a sequence of polynomials , 2001 .

[31]  A. Seidenberg.  A NEW DECISION METHOD FOR ELEMENTARY ALGEBRA , 1954 .

[32]  A. A. Markov.  On the Inversion Complexity of a System of Functions , 1958, JACM.

[33]  Dima Grigoriev,et al.  Solving Systems of Polynomial Inequalities in Subexponential Time , 1988, J. Symb. Comput..

[34]  Marie-Françoise Roy,et al.  On the combinatorial and algebraic complexity of Quanti erEliminationS , 1994 .

[35]  R. Impagliazzo,et al.  Subexponential Circuits : Derandomizing the XOR Lemma , 2003 .

[36]  S. Shelah.  A combinatorial problem; stability and order for models and theories in infinitary languages. , 1972 .

[37]  Marek Karpinski,et al.  Randomized ( n 2 ) Lower Bound for , 2007 .

[38]  John F. Canny,et al.  Computing Roadmaps of General Semi-Algebraic Sets , 1991, Comput. J..

[39]  On Randomized Semi-algebraic Test Complexity , 1993, J. Complex..

[40]  Joos Heintz,et al.  Sur la complexité du principe de Tarski-Seidenberg , 1989 .

[41]  Marek Karpinski,et al.  On real Turing machines that toss coins , 1995, STOC '95.

[42]  Vladimir Vapnik,et al.  Chervonenkis: On the uniform convergence of relative frequencies of events to their probabilities , 1971 .

[43]  J. Renegar,et al.  On the Computational Complexity and Geometry of the First-Order Theory of the Reals, Part I , 1989 .

[44]  Rusins Freivalds,et al.  Probabilistic Machines Can Use Less Running Time , 1977, IFIP Congress.

[45]  James H. Davenport,et al.  Real Quantifier Elimination is Doubly Exponential , 1988, J. Symb. Comput..