Title:
METHOD FOR DYNAMICALLY GENERATING ADDITIONAL TERMS FOR EACH MEANING OF EVERY NATURAL LANGUAGE EXPRESSION; DICTIONARY MANAGER, DOCUMENTATION GENERATOR, TERM ANNOTATOR, SEARCH SYSTEM, AND DEVICE FOR BUILDING DOCUMENT INFORMATION SYSTEM BASED ON THE METHOD
Document Type and Number:
WIPO Patent Application WO/2011/155736
Kind Code:
A9
Abstract:
The present invention relates to changing an information system comprising natural language expressions to an information system based on unit expressions of meaning, which is accompanied by functional changes for an information search system, term dictionary, documentation generator, and term converter. The accuracy of current search systems is very low. This is because natural language represents many meanings using few words. Due to the problem of expressions becoming longer and more difficult to recollect as the number of terms increases, people use a small number of terms in a repetitive manner. When unit expressions of meaning having 1 term corresponding to 1 meaning are introduced, the accuracy of a search system can approach 100%. The present invention presents a method for easily generating unit expressions of meaning, and a method for efficiently applying the generated unit expressions of meaning to documents from around the world. The method for creating unit expressions of meaning is a technique of breaking down each natural language term into the number of respective meanings thereof. Because this is a matter of a simple breakdown of terms, anyone can generate expressions. The task of applying generated terms to documents from around the world is formidable. For this task, according to the present invention, instead of changing each word that is repetitively used, alignment is performed for each word, and certain aligned word groups are simultaneously processed. Even if one word has been used several hundred billion times in documents throughout the world, there is no need to perform term conversions several hundred billion times. If the word in question has several meanings, the task of conversion can be performed simply by way of several sorting commands. Even if the repetitive use of terms does not impose a large load on term conversion, because the number of unit expressions of meaning itself is enormous, term conversion is not simple. The task of processing close to 10 billion unit expressions of meaning is daunting. A method for solving this difficulty is to equally distribute the task to a number of users. The greatest factor contributing to the ambiguity of natural language is the presence of innumerable proper nouns. These encroach on the domains of nouns, adjectives, verbs, and all other parts of speech, causing semantic confusion. While not limited to people's names, when considering proper nouns only in that context, there are over 10 billion terms in this category since the global population exceeds 6 billion. The present invention persentss a configuration in which this prodigious task is equally allotted to a countless number of users. Users having needs may perform tasks to fulfill their requirements and benefit from their work. If the users feel that term conversion is required, the users may perform term generation and term conversion tasks so that a state that is always satisfactory for users can be maintained. The present invention provides: 1) a unit expression of meaning dictionary manager that can easily generate unit expressions of meaning; and 2) a search annotator which is a means for categorizing words and converting (annotating) words belonging to a word group into unit expressions of meaning. The annotator operates as part of a search system. The alignment and search of words uses existing search system functions. Also the present invention provides 3) a unit expression of meaning converter (annotator) performing a function similar to the search annotator. The task of making a global information system based on unit expressions of meaning is an enormous endeavor. However, the problem of natural language being unclear in meaning presents a large obstacle for development in many fields. The present invention provides a basis for achieving considerable advances in the semantic web field, search system field, language translation field, and artificial intelligence field, by providing clear language thereto.
Inventors:
PARK DONG MIN (KR)
Application Number:
PCT/KR2011/004113
Publication Date:
June 21, 2012
Filing Date:
June 06, 2011
Export Citation:
Assignee:
PARK DONG MIN (KR)
International Classes:
G06F17/21; G06F17/30; G06F40/00
Download PDF:
Claims: