Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Statistical method for taxonomy oriented dictionary design and machine translation production system for multi-taxonomy documents

IP.com Disclosure Number: IPCOM000018980D
Original Publication Date: 2003-Aug-25
Included in the Prior Art Database: 2003-Aug-25
Document File: 3 page(s) / 44K

Publishing Venue

IBM

Abstract

Disclosed is a program for enabling to select the most appropriate Japanese equivalent word in machine translations.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 3

  Statistical method for taxonomy oriented dictionary design and machine translation production system for multi-taxonomy documents

Calculate "Taxonomy Vector" of a source text First, calculate "Taxonomy Vector" of a source text. Taxonomy Vector represents what the text relates with in vector style. For example the element of the vector is Computer, Biology, Economy, Literature etc. If a source text describes economy, the vector element of Economy is large. Taxonomy Vector is calculated by what kind of categorizer.

Source text Categorizer Taxonomy

Vector

Calculate similarity level between the source text and dictionaries

Very many translation dictionaries are prepared. All of them have unique Taxonomy

  Vector. Calculate similarity level between the source text and dictionaries using Taxonomy

  Vector of the dictionaries and the source text. Any kind of methods to calculate similarity of two vectors are acceptable.

Dictionary

Dictionary

Dictionary

Dictionary

Dictionary

Translation

Translate the source text into a target language. Translation engine prioritizes the dictionary with the Taxonomy Vector which is similar to the vector of the source text.

1

Taxonomy Vector

[This page contains 3 pictures or other non-text objects]

Page 2 of 3

Dictionaries' Taxonomy Vector

Source text's Taxonomy Vector

Primary dictionary's vector

Secondary dictionary's vector

Correct the result of translation

Correct the result of translation if there are invalid translations and unknown words. Based on these corrections, create new dictionary. Optionally, the words which are not corrected can be added new dictionary as...