Data processing: speech signal processing – linguistics – language – Linguistics – Translation machine
Reexamination Certificate
1998-01-07
2004-09-07
Edouard, Patrick N. (Department: 2654)
Data processing: speech signal processing, linguistics, language
Linguistics
Translation machine
C704S007000
Reexamination Certificate
active
06789057
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method of managing bilingual dictionaries distributed on a network and a machine translation method using such method.
2. Description of the Related Art
U.S. Pat. No. 4,641,264, JP-A-6-259600, JP-A-5120336 and U.S. Pat. No. 4,502,128 describe a dictionary retrieval, a translation knowledge, a URL (uniform resource location), an HTML (hypertext markup language) and a dictionary distribution server/client system.
At present, in accordance with a development of a network, in particular, a widespread of an internet, users are able to easily access information of foreign languages which comprise mainly English.
The JP-A-6-259600, for example, describes an example in which users share a machine translation dictionary by effectively utilizing a network, and proposes a system which is connected to the network via dictionaries. When this system is applied to a translation, translation dictionaries of respective languages such as English and Russian are previously connected to the system and a language of an inputted word is discriminated. Thereafter, inputted data (image data of character) is transferred to a center having a corresponding translation dictionary. According to the above-mentioned related art, dictionaries that are used to translate specific fields or specific languages are limited, and words that could not be translated such dictionaries are processed as words which cannot be translated.
In the machine translation, a quality of a dictionary and a coverage of words considerably affect the quality of the translation. However, information on the internet covers a considerably wide field. Moreover, new information appears on the internet day by day so that it takes the high cost for supplement dictionaries in corresponding to the increase of words of a wide field and new words.
Further, service providers or makers of machine translation machines having dictionaries supplemented by other servers are used to provide dictionaries of new versions at every several months. There is then the problem that dictionary information necessary for users can not be offered to the users timely.
In the translation processing, since users access a dictionary frequently, a dictionary access speed considerably affects a processing speed. In the internet, it cannot be expected that its line speed is higher than that of a LAN (local area network) so that, when the user accesses a plurality of dictionaries via the internet during the translation processing, the translation cannot be completed at high speed. Furthermore, at present to connect a client computer installed on a home user to the internet, it is necessary to use a line such as a telephone line of which the line speed is very low as compared with that of the LAN. As a consequence, it takes a lot of time for the home user to access the dictionaries through the network, which is not suitable in actual practice.
SUMMARY OF THE INVENTION
In view of the aforesaid aspect, it is an object of the present invention to provide a dictionary server in which a cost for retrieving desired dictionary information can be suppressed by executing retrieval requests in the sequential order of high priorities given to other dictionary servers having dictionary information of a dictionary entry or other dictionary servers having a large possibility that they will have dictionary information of a dictionary entry if the dictionary entry of retrieval request does not exist on the dictionary server at the time a dictionary information retrieval request is issued.
It is another object of the present invention to provide a recording medium for use with a general information processing apparatus and which can be read by a computer system which stores therein processing procedures (a set of programs and command codes) and necessary data for translating a natural language into other natural language in a machine translation fashion.
It is a further object of the present invention to provide a machine translation method in which a translated sentence of a high quality can be created by retrieving dictionary information through a network even if a processing speed of a dictionary retrieval through the network is not sufficient when a client executes a translation processing.
It is yet a further object of the present invention to provide a dictionary information collection method for a dictionary server in which a dictionary constructing work is supported by preparing an unknown word having a large possibility that it will be requested to be retrieved before a client issues a dictionary information retrieval request in the dictionary server so that a responsiveness obtained at the time the dictionary information retrieval request is issued can be made at a high speed.
According to the present invention, there is provided a machine translation method of translating a document by using a translation knowledge. The machine translation method comprises the steps of transmitting an inquiry for inquiring a translation knowledge used to translate a syntactic unit containing an unknown word and having a correct grammar to a network when an unknown word that cannot be translated is recognized during a document is translated, continuing translating a syntax following the syntactic unit, and completing the translation of the syntactic unit by using an answer when the answer to the inquiry is received.
The inquiry step may include a first sub-step of outputting the syntactic unit translated in the form of having the unknown word temporarily processed when the unknown word that cannot be translated is recognized, and the syntax completion step may include a second sub-step of outputting the syntactic unit in which the unknown word in the syntactic unit is translated by using the received answer.
The first sub-step may include a sub-step of translating the syntactic unit having the correct grammar into the syntactic unit translated in the form of having the unknown word temporarily processed by estimating a part of speech to which the unknown word belongs. The second step may include a sub-step of comparing the estimated part of speech with a part of speech of the unknown word contained in the answer, translating the unknown word if they agree with each other, and translating again the syntactic unit based on the answer if they do not agree with each other.
According to the present invention, it is possible to provide a retrieval method of retrieving a dictionary stored in a storage in response to a request in an information processing apparatus connected to a network. The retrieval method comprises the steps of retrieving the dictionary of the information processing apparatus when a dictionary retrieval request concerning a dictionary entry of a document is received and transmitting a retrieval request from the information processing apparatus to the network when the dictionary entry does not exist in the dictionary. The transmission step may include a step of determining a transmission destination to which the retrieval request is transmitted based on priorities supplied to a plurality of other dictionary servers by retrieval results received to a plurality of past retrieval requests and a field of the dictionary entry.
In order to construct a dictionary, there may be executed a step of determining importance of a plurality of documents based on respective frequencies of retrieval requests to a plurality of documents on a network, a step of retrieving an important document from a plurality of documents in order to recognize an unknown word, a step of outputting an important unknown word to an output apparatus together with an input request of a corresponding equivalent and a step of memorizing the equivalent inputted on the output apparatus in the dictionary.
Furthermore, according to the present invention, it is possible to provide a computer readable storage medium for use with an information processing apparatus connected to a network for translating a document by using a translation knowle
Matsuda Jun-ichi
Morimoto Yasutsugu
Antonelli Terry Stout & Kraus LLP
Edouard Patrick N.
LandOfFree
Dictionary management method and apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Dictionary management method and apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dictionary management method and apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3189191