DocumentCode :
1780744
Title :
A Composition Theorem for Parity Kill Number
Author :
O´Donnell, Ryan ; Wright, John ; Yu Zhao ; Xiaorui Sun ; Li-Yang Tan
Author_Institution :
Comput. Sci. Dept., Carnegie Mellon Univ., Pittsburgh, PA, USA
fYear :
2014
fDate :
11-13 June 2014
Firstpage :
144
Lastpage :
154
Abstract :
In this work, we study the parity complexity measures Cmin[f] and DT[f]. Cmin[f] is the parity kill number of f, the fewest number of parities on the input variables one has to fix in order to "kill" f, i.e. To make it constant. DT[f] is the depth of the shortest parity decision tree which computes f. These complexity measures have in recent years become increasingly important in the fields of communication complexity [1], [2], [3], [4] and pseudorandomness [5], [6], [7]. Our main result is a composition theorem for Cmin. The k-th power of f, denoted fok, is the function which results from composing f with itself k times. We prove that if f is not a parity function, then Cmin[fok] ≥ Ω (Cmin[f]k). In other words, the parity kill number of f is essentially super multiplicative in the normal kill number of f (also known as the minimum certificate complexity). As an application of our composition theorem, we show lower bounds on the parity complexity measures of Sortok and HIok. Here sort is the sort function due to Ambainis [8], and HI is Kushilevitz\´s hemi-icosahedron function [9]. In doing so, we disprove a conjecture of Montanaro and Osborne [2] which had applications to communication complexity and computational learning theory. In addition, we give new lower bounds for conjectures of [2], [3] and [4].
Keywords :
communication complexity; decision trees; Kushilevitz hemi-icosahedron function; communication complexity; composition theorem; computational learning theory; lower bounds; minimum certificate complexity; parity complexity measures; parity kill number; pseudorandomness; shortest parity decision tree; sort function; Boolean functions; Complexity theory; Computer science; Decision trees; Educational institutions; Handheld computers; Input variables; communication complexity; log-rank conjecture; parity decision trees;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computational Complexity (CCC), 2014 IEEE 29th Conference on
Conference_Location :
Vancouver, BC
Type :
conf
DOI :
10.1109/CCC.2014.22
Filename :
6875483
Link To Document :
بازگشت