{"title":"Fast and efficient construction of an unbiased random sequence","authors":"B. Ryabko, E. Matchikina","doi":"10.1109/ISIT.1998.709077","DOIUrl":null,"url":null,"abstract":"The problem of efficiently converting an input sequence of symbols generated by a stationary random process into a sequence of independent, equiprobable output symbols was considered by von Neumann. Elias (1972) proposed a method in which the sequence of symbols was divided into blocks of length N. Elias' method requires one to store all 2/sup N/ code words. That is why the memory size of the encoder increases exponentially when N grows. We suggest a new method of constructing an unbiased random sequence which requires the memory size of O(Nlog/sup 2/N) bits and has a time of calculation O(log/sup 3/Nloglog(N)) bit operations per letter as N/spl rarr//spl infin/.","PeriodicalId":133728,"journal":{"name":"Proceedings. 1998 IEEE International Symposium on Information Theory (Cat. No.98CH36252)","volume":"146 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1998-08-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"36","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings. 1998 IEEE International Symposium on Information Theory (Cat. No.98CH36252)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISIT.1998.709077","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 36
Abstract
The problem of efficiently converting an input sequence of symbols generated by a stationary random process into a sequence of independent, equiprobable output symbols was considered by von Neumann. Elias (1972) proposed a method in which the sequence of symbols was divided into blocks of length N. Elias' method requires one to store all 2/sup N/ code words. That is why the memory size of the encoder increases exponentially when N grows. We suggest a new method of constructing an unbiased random sequence which requires the memory size of O(Nlog/sup 2/N) bits and has a time of calculation O(log/sup 3/Nloglog(N)) bit operations per letter as N/spl rarr//spl infin/.