English

Improved Deterministic Length Reduction

Data Structures and Algorithms 2008-02-04 v1

Abstract

This paper presents a new technique for deterministic length reduction. This technique improves the running time of the algorithm presented in \cite{LR07} for performing fast convolution in sparse data. While the regular fast convolution of vectors V1,V2V_1,V_2 whose sizes are N1,N2N_1,N_2 respectively, takes O(N1logN2)O(N_1 \log N_2) using FFT, using the new technique for length reduction, the algorithm proposed in \cite{LR07} performs the convolution in O(n1log3n1)O(n_1 \log^3 n_1), where n1n_1 is the number of non-zero values in V1V_1. The algorithm assumes that V1V_1 is given in advance, and V2V_2 is given in running time. The novel technique presented in this paper improves the convolution time to O(n1log2n1)O(n_1 \log^2 n_1) {\sl deterministically}, which equals the best running time given achieved by a {\sl randomized} algorithm. The preprocessing time of the new technique remains the same as the preprocessing time of \cite{LR07}, which is O(n12)O(n_1^2). This assumes and deals the case where N1N_1 is polynomial in n1n_1. In the case where N1N_1 is exponential in n1n_1, a reduction to a polynomial case can be used. In this paper we also improve the preprocessing time of this reduction from O(n14)O(n_1^4) to O(n13polylog(n1))O(n_1^3{\rm polylog}(n_1)).

Keywords

Cite

@article{arxiv.0802.0017,
  title  = {Improved Deterministic Length Reduction},
  author = {Amihood Amir and Klim Efremenko and Oren Kapah and Ely Porat and Amir Rothschild},
  journal= {arXiv preprint arXiv:0802.0017},
  year   = {2008}
}

Comments

7 pages

R2 v1 2026-06-21T10:08:29.573Z