A probabilistic model of computing with words

  • Authors:

  • Daowen Qiu;Huaiqing Wang

  • Affiliations:

  • Department of Computer Science, Zhongshan University, Guangzhou 510275, P. R. China and State Key Laboratory of Intelligent Technology and Systems, Department of Computer Science and Technology, T ...;Department of Information Systems, City University of Hong Kong, Hong Kong

  • Venue:
  • Journal of Computer and System Sciences
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Computing in the traditional sense involves inputs with strings of numbers and symbols rather than words, where words mean probability distributions over input alphabet, and are different from the words in classical formal languages and automata theory. In this paper our goal is to deal with probabilistic finite automata (PFAs), probabilistic Turing machines (PTMs), and probabilistic context-free grammars (PCFGs) by inputting strings of words (probability distributions). Specifically, (i) we verify that PFAs computing strings of words can be implemented by means of calculating strings of symbols (Theorem 1); (ii) we elaborate on PTMs with input strings of words, and particularly demonstrate by describing Example 2 that PTMs computing strings of words may not be directly performed through only computing strings of symbols, i.e., Theorem 1 may not hold for PTMs; (iii) we study PCFGs and thus PRGs with input strings of words, and prove that Theorem 1 does hold for PCFRs and PRGs (Theorem 2); a characterization of PRGs in terms of PFAs, and the equivalence between PCFGs and their Chomsky and Greibach normal forms, in the sense that the inputs are strings of words, are also presented. Finally, the main results obtained are summarized, and a number of related issues for further study are raised.