A Natural Language Model of Computing with Words in Web Pages

A Natural Language Model of Computing with Words in Web Pages Zheng Ze-yu1, Zhang Ping2 1School of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan, Hubei, 430074, P.R. China 2School of Foreign Languages, Zhongnan University of Economics and Law, Wuhan, Hubei 430064, P.R. China, seekinghome@126.com Abstract: This paper is to deal with probabilistic finite automata (PFAs) by inputting strings of words (probability distributions) in web pages of Internet, and investigates the development of a natural language (NL) interface for mixed initiative dialogues within a constrained domain and demonstrates the applicability. Specifically, it is verified that PFAs computing strings of words can be implemented by means of calculating strings of symbols and a characterization of PFAs. On the other hand, a natural language model system in Internet is discussed. The system is written entirely in a purely functional language, which results in a surprisingly small and simple program. Finally, the main results obtained are summarized, and a number of related issues for further study are raised. Keywords: Natural language processing, Functional programming, Web pages, Probabilistic finite automata 1 Introduction With the deregulation of the Internet and the development of ever more complex and sophisticated systems such as enterprise resource planning, user relationship management systems, and operating systems it was inevitable that a parallel requirement for more sophisticated online user assistance systems would follow. Computing in the traditional sense involves inputs with strings of numbers and symbols rather than words, and even in those nontraditional models of computation such as molecular, stochastic, analog, and quantum computing, the inputs still are strings of numbers and symbols instead of words. In this paper, words mean probability distributions over input alphabet, and are different from the words in classical formal languages and automata theory that represent strings of input symbols. The major technical contributions of this paper are twofold. On one hand, we study PFAs with the inputs to be strings of words that are probability distributions over input alphabets[1]. On the other hand, we do research on the natural language system which contains the natural language subsystem that comprises a parser that interprets natural language inputs to semantic information and the knowledge-based subsystem that selects the most appropriate script for a guided dialogue to the user with semantic tags extracted by the natural language subsystem. In the paper, we deal with PFAs with input strings of words. Specifically, we can derive the probability for PFAs computing strings of words by means of calculating these probabilities of PFAs recognizing strings of symbols (Theorem 1), and thus shows that PFAs for inputting strings of words can be transferred to PFAs computing the convention all strings of symbols, at the expense of more steps of computation, which may be called computation tractability. This paper builds upon the previous research to present a new two-part computational natural language processing (NLP) system based upon the use of an executable set of functional equations and through this notation demonstrates its applicability to the creation of knowledge-based online …

Content truncated to 500 words