Computing with unreliable information

Uriel Feige; David Peleg; Prabhakar Raghavan; Eli Upfal

doi:10.1145/100216.100230

Computing with unreliable information

Uriel Feige, David Peleg, Prabhakar Raghavan, Eli Upfal

Weizmann Institute of Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

70 Scopus citations

Abstract

A (possibly randomized) computation tree is considered in which each node gives the correct answer with some probability ≥ p, where p is a fixed constant in ( 1/2 , 1), bounded away from 1/2 and 1. The node faults are independent. The depth of the computation tree is studied in terms of a tolerance parameter Q member of (0, 1/2 ): on any instance, the computation tree leads to a leaf giving the correct answer on that instance with probability at least 1 - Q. The success probability of the algorithm is computed over the combined probability space of the outcome of individual comparisons and the results of coinflips (in case our algorithm is randomized). Noisy comparison trees are studied for problems such as sorting, selection and searching.

Original language	English
Title of host publication	Proc 22nd Annu ACM Symp Theory Comput
Publisher	Publ by ACM
Pages	128-137
Number of pages	10
ISBN (Print)	0897913612, 9780897913614
DOIs	https://doi.org/10.1145/100216.100230
State	Published - 1990
Externally published	Yes
Event	Proceedings of the 22nd Annual ACM Symposium on Theory of Computing - Baltimore, MD, USA Duration: 14 May 1990 → 16 May 1990

Publication series

Name	Proc 22nd Annu ACM Symp Theory Comput

Conference

Conference	Proceedings of the 22nd Annual ACM Symposium on Theory of Computing
City	Baltimore, MD, USA
Period	14/05/90 → 16/05/90

Access to Document

10.1145/100216.100230

Cite this

@inproceedings{d64c649b8ffa4d78b698379fa892c1b7,

title = "Computing with unreliable information",

abstract = "A (possibly randomized) computation tree is considered in which each node gives the correct answer with some probability ≥ p, where p is a fixed constant in ( 1/2 , 1), bounded away from 1/2 and 1. The node faults are independent. The depth of the computation tree is studied in terms of a tolerance parameter Q member of (0, 1/2 ): on any instance, the computation tree leads to a leaf giving the correct answer on that instance with probability at least 1 - Q. The success probability of the algorithm is computed over the combined probability space of the outcome of individual comparisons and the results of coinflips (in case our algorithm is randomized). Noisy comparison trees are studied for problems such as sorting, selection and searching.",

author = "Uriel Feige and David Peleg and Prabhakar Raghavan and Eli Upfal",

year = "1990",

doi = "10.1145/100216.100230",

language = "אנגלית",

isbn = "0897913612",

series = "Proc 22nd Annu ACM Symp Theory Comput",

publisher = "Publ by ACM",

pages = "128--137",

booktitle = "Proc 22nd Annu ACM Symp Theory Comput",

note = "Proceedings of the 22nd Annual ACM Symposium on Theory of Computing ; Conference date: 14-05-1990 Through 16-05-1990",

}

TY - GEN

T1 - Computing with unreliable information

AU - Feige, Uriel

AU - Peleg, David

AU - Raghavan, Prabhakar

AU - Upfal, Eli

PY - 1990

Y1 - 1990

N2 - A (possibly randomized) computation tree is considered in which each node gives the correct answer with some probability ≥ p, where p is a fixed constant in ( 1/2 , 1), bounded away from 1/2 and 1. The node faults are independent. The depth of the computation tree is studied in terms of a tolerance parameter Q member of (0, 1/2 ): on any instance, the computation tree leads to a leaf giving the correct answer on that instance with probability at least 1 - Q. The success probability of the algorithm is computed over the combined probability space of the outcome of individual comparisons and the results of coinflips (in case our algorithm is randomized). Noisy comparison trees are studied for problems such as sorting, selection and searching.

AB - A (possibly randomized) computation tree is considered in which each node gives the correct answer with some probability ≥ p, where p is a fixed constant in ( 1/2 , 1), bounded away from 1/2 and 1. The node faults are independent. The depth of the computation tree is studied in terms of a tolerance parameter Q member of (0, 1/2 ): on any instance, the computation tree leads to a leaf giving the correct answer on that instance with probability at least 1 - Q. The success probability of the algorithm is computed over the combined probability space of the outcome of individual comparisons and the results of coinflips (in case our algorithm is randomized). Noisy comparison trees are studied for problems such as sorting, selection and searching.

UR - http://www.scopus.com/inward/record.url?scp=0025065888&partnerID=8YFLogxK

U2 - 10.1145/100216.100230

DO - 10.1145/100216.100230

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:0025065888

SN - 0897913612

SN - 9780897913614

T3 - Proc 22nd Annu ACM Symp Theory Comput

SP - 128

EP - 137

BT - Proc 22nd Annu ACM Symp Theory Comput

PB - Publ by ACM

T2 - Proceedings of the 22nd Annual ACM Symposium on Theory of Computing

Y2 - 14 May 1990 through 16 May 1990

ER -

Computing with unreliable information

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this