Amplification by Read-Once Formulas

  title={Amplification by Read-Once Formulas},
  author={Moshe Dubiner and Uri Zwick},
  journal={SIAM J. Comput.},
Moore and Shannon have shown that relays with arbitrarily high reliability can be built from relays with arbitrarily poor reliability. Valiant used similar methods to construct monotone read-once formulas of size $O(n^{\alpha+2})$ (where $\alpha=\log_{\sqrt{5}-1}2\simeq 3.27$) that amplify $(\psi-\frac{1}{n},\psi+\frac{1}{n})$ (where $\psi=(\sqrt{5}-1)/2\simeq0.62$) to $(2^{-n},1-2^{-n})$ and deduced as a consequence the existence of monotone formulas of the same size that compute the majority… 

Figures from this paper

Tight Space Complexity of the Coin Problem

A tight threshold is shown between the values of $\beta=n^{-c}$ where $O(\log n) $ width suffices and the regime where $poly (n)$ width is needed, with a transition at $c=1/3$.

Monotone Circuits for the Majority Function

It is shown that the size can be further reduced at the expense of increased depth, and the circuit obtained is for the majority of size and depth about $n^{1+\sqrt{2}}$ and 9.9logn.

Growth processes on formulas and reversible circuits

This thesis parameterizes a growth process by its initial conditions and characterize it by the existence and shape of a limiting probability distribution that describes the likelihood of it realizing a particular Boolean function, and proposes and characterize growth processes on reversible circuits.


Two new complexity measures for Boolean functions are introduced, which are named sumPI and maxPI, and the main result is proven via a combinatorial lemma which relates the square of the spectral norm of a matrix to the squares ofthe spectral norms of its submatrices.

On the Size of Probabilistic Formulae

  • H. Klauck
  • Mathematics, Computer Science
  • 1997
A lower bound method for probabilistic formulae based on the VC-dimension and the Neciporuk lower bound is devised and lower bounds on the size of 2-sided bounded error, Monte Carlo, and Las Vegas formULae over an arbitrary basis are given.

The Boolean functions computed by random Boolean formulas or how to grow the right function

Growth processes are characterized by their initial conditions to derive conditions under which results such as Valiant's J Algorithms 5 (1984), 363–366 hold and explicit bounds on the convergence rates of several growth processes are obtained.

On linear rewriting systems for Boolean logic and some applications to proof theory

This paper studies properties of systems consisting only of linear inferences and finds that the length of any 'nontrivial' derivation in such a system is bound by a polynomial, unless coNP=NP.

Balanced And/Or Trees and Linear Threshold Functions

It is proved that the probability distribution induced on Boolean functions is shown to have a limit when letting the depth of these formulas grow to infinity, and that its support is made of linear threshold functions.

Analysis of random processes via And-Or tree evaluation

These tools provide a unifying, intuitive, and powerful framework for carrying out the analysis of several previously studied random processes of interest, including random loss-resilient codes, solving random k-SAT formula using the pure literal rule, and the greedy algorithm for matchings in random graphs.

No complete linear term rewriting system for propositional logic

It is shown that, as long as reduction steps are polynomial-time decidable, such a rewriting system does not exist unless coNP=NP, i.e. that every Boolean tautology can be written as a (left- and right-) linear rewrite rule.



Amplification of Bounded Depth Monotone Read-Once Boolean Formulae

The size of monotone read-once formulae in $\Sigma _d \cup \Pi _d $ that amplify $(p,p + 1 / m)$ to $(p',p' + 1/ c)$ is $\exp (\theta ((d - m)(m / c)^{1 / (d - 1)} ))$ under certain conditions.

Relative to a Random Oracle A, PA != NPA != co-NPA with Probability 1

Let A be a language chosen randomly by tossing a fair coin for each string x to determine whether x belongs to A, and${\bf NP}^A is shown, with probability 1, to contain a-immune set, i.e., a set having no infinite subset in ${\bf P]^A $.

Complexity of the realization of a linear function in the class of II-circuits

AbstractIt is proved that the linear function gn(x1,..., xn) = x1 + ... + xnmod 2 is realized in the class of II-circuits with complexity Lπ(gn) ≥n2. Combination of this result with S. V.

A theorem on probabilistic constant depth Computations

Stockmeyer [St] showed that probabilistic bounded depth circuits can approximate the exact number of ones in the input with very low probability of error.

Faster circuits and shorter formulae for multiple addition, multiplication and symmetric Boolean functions

The shallowest known multiplication circuits are constructed on the basis of the optimal constructions of multiple carry-save adders using any given basic addition unit.

Two theorems on random polynomial time

  • L. Adleman
  • Computer Science
    19th Annual Symposium on Foundations of Computer Science (sfcs 1978)
  • 1978
Where the traditional method of polynomial reduction has been inapplicable, randomness has been used in demonstrating intractibility by Adleman and Manders, and in showing problems equivalent by Rabin, a new examination of randomness is in order.

Optimal carry save networks

In this paper simple basic carry save adders are described using which multiplication circuits of depth 3.71 log n and majority formulae of size O (n3.13) are constructed and the shallowest known multiplication circuits and the shortest formULae for the majority function are obtained.

Amplification of probabilistic boolean formulas

  • R. Boppana
  • Computer Science
    26th Annual Symposium on Foundations of Computer Science (sfcs 1985)
  • 1985
This paper shows that the amount of amplification that Valiant obtained is optimal, and gives an O(k4.3 n log n) upper bound for the size of monotone formulas computing the kth threshold function of n variables.

On the Shrinkage Exponent for Read-Once Formulae

How Do Read-Once Formulae Shrink?

It is shown that f e depends, on the average, on only O (e α n + e n 1/α ) variables, where .