Automated system for generating natural language translations that are domain-specific, grammar rule-based, and/or based on part-of-speech analysis

• US 6,278,967 B1
• Filed: 12/16/1998
• Issued: 08/21/2001
• Est. Priority Date: 08/31/1992
• Status: Expired due to Term

• Current Assignee: Cantena Service Agent Corporation
• Original Assignee: Logovista Corporation (Wayo Kosan Co., Ltd.)
• Inventors: Kuno, Susumu, Akers, Glenn A.
• Primary Examiner(s): Thomas, Joseph
• Application Number:
  US09/171,242
• Time in Patent Office: 979 Days
• Field of Search: 704/2, 704/3, 704/4, 704/5, 704/6, 704/7, 704/8, 704/9, 704/10, 704/1, 707/536, 707/532, 707/531
• US Class Current: 704/2
• CPC Class Codes: G06F 40/211 Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models
• CPC Class Codes: G06F 40/55 Rule-based translation

Abstract

Systems and methods for automated natural language translation take source natural language and translate it into a target natural language. The language translation is facilitated through the use of domain specific dictionaries, controlling grammar rules, and probability values associated with some dictionary entries. Domain specific dictionaries provide different translations of words depending on the context in which the words are used. Controlling grammar rules are used to supersede all other grammar rules to eliminate inaccurate and undesirable translations. Probability values are used to determine the likelihood a particular multi-word combination is a particular part of speech.

Claims

1. An automated natural language translation system, comprising:

computer storage means;

means for receiving input textual information in a source natural language and for storing the input textual information in the computer storage means;

a plurality of keywords for at least one domain, the keywords for each domain being stored in the computer storage means;

a dictionary including a plurality of words wherein at least some of the words have translations in a target natural language which are different depending on the domain in which those words are used, the dictionary being stored in the computer storage means; and

a translation engine for accessing the computer storage means and for translating the input textual information in the source natural language into output textual information in the target natural language, the translation engine creating and using a tree structure in performing the translation, the translation engine determining at least one domain to which at least a portion of the input textual information belongs by addressing each sentence in the input textual information one at a time and altering the probability of certain words in a sentence being translated into the target natural language with a translation specific to a particular domain based on at least one of the keywords.

2. An automated natural language translation system, comprising:

computer storage means;

means for receiving input textual information in a source natural language and for storing the input textual information in the computer storage means;

a plurality of grammar rules stored in the computer storage means, at least one of the grammar rules being designated as a controlling grammar rule which blocks all of the other grammar rules for any portion of the input textual information to which the controlling grammar rule applies; and

a translation engine for accessing the computer storage means and for translating the input textual information in the source natural language into output textual information in a target natural language, the translation engine creating and using a tree structure in performing the translation, the translation engine including a parser for applying the grammar rules to the input textual information and for ignoring all other grammar rules for a portion of the input textual information when the controlling grammar rule applies to that portion of the input textual information.

3. An automated natural language translation system, comprising:

computer storage means;

means for receiving input textual information in a source natural language and for storing the input textual information in the computer storage means;

a plurality of grammar rules stored in the computer storage means, each of at least some of the grammar rules being associated with one of a plurality of predetermined parts of speech (POS), at least one of the grammar rules associated with one of the predetermined POS being designated as a controlling grammar rule which blocks all of the other grammar rules associated with that predetermined POS for any portion of the input textual information to which the controlling grammar rule applies, and

a translation engine for accessing the computer storage means and for translating the input textual information in the source natural language into output textual information in a target natural language, the translation engine creating and using a tree structure in performing the translation, the translation engine including a parser for applying the grammar rules to the input textual information and for ignoring all other grammar rules associated with one of the predetermined POS for a portion of the input textual information when the controlling grammar rule associated with that predetermined POS applies to that portion of the input textual information.

4. An automated natural language translation system, comprising:

computer storage means;

means for receiving input textual information in a source natural language and for storing the input textual information in the computer storage means;

a dictionary including a plurality of multi-word entries wherein at least one of the multi-word entries being a marked entry which has associated therewith a value indicating the likelihood the marked entry will be translated into a target natural language as a particular part of speech, the dictionary and the values being stored in the computer storage means; and

a translation engine for accessing the computer storage means and for translating the input textual information in the source natural language into output textual information in the target natural language, the translation engine creating and using a tree structure in performing the translation, the translation engine including a parser which is more or less likely to parse the marked entry appearing in a sentence of the input textual information as the particular part of speech depending on the value associated with the marked entry.