WO2008029881A1 - Système de traitement du langage naturel et système d'enregistrement de dictionnaire - Google Patents
Système de traitement du langage naturel et système d'enregistrement de dictionnaire Download PDFInfo
- Publication number
- WO2008029881A1 WO2008029881A1 PCT/JP2007/067416 JP2007067416W WO2008029881A1 WO 2008029881 A1 WO2008029881 A1 WO 2008029881A1 JP 2007067416 W JP2007067416 W JP 2007067416W WO 2008029881 A1 WO2008029881 A1 WO 2008029881A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- dictionary
- data
- input
- dictionary data
- registration
- Prior art date
Links
- 238000003058 natural language processing Methods 0.000 title claims abstract description 122
- 238000000034 method Methods 0.000 claims abstract description 88
- 238000012545 processing Methods 0.000 claims description 49
- 238000013519 translation Methods 0.000 claims description 26
- 230000006870 function Effects 0.000 claims description 25
- 230000015572 biosynthetic process Effects 0.000 claims description 14
- 238000006243 chemical reaction Methods 0.000 claims description 14
- 238000003786 synthesis reaction Methods 0.000 claims description 14
- 238000001514 detection method Methods 0.000 claims description 11
- 240000000220 Panda oleosa Species 0.000 claims description 9
- 235000016496 Panda oleosa Nutrition 0.000 claims description 9
- 230000000694 effects Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 108090000623 proteins and genes Proteins 0.000 description 6
- 238000004590 computer program Methods 0.000 description 2
- 238000012790 confirmation Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000005236 sound signal Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L13/00—Speech synthesis; Text to speech systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/10—Text processing
- G06F40/12—Use of codes for handling textual entities
- G06F40/126—Character encoding
- G06F40/129—Handling non-Latin characters, e.g. kana-to-kanji conversion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/242—Dictionaries
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/247—Thesauruses; Synonyms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/40—Processing or translation of natural language
- G06F40/42—Data-driven translation
- G06F40/49—Data-driven translation using very large corpora, e.g. the web
Definitions
- the present invention relates to a technology for processing a natural language using language information stored in a dictionary, and in particular, a natural language processing system capable of easily updating a dictionary, and an application for the system.
- the present invention relates to a dictionary registration system.
- Kana-Kanji conversion, machine translation, speech recognition, speech synthesis, and! / The natural language processing system is basically based on the words stored in the dictionary and the language information assigned to them! It is difficult to appropriately process words that are processed and are not stored in the dictionary, that is, unknown words. On the other hand, natural languages are changing day by day, and new words are born, and even if they are known words, new usages occur, so it is difficult to store them all in the dictionary in advance. Therefore, this problem has been addressed by providing a user dictionary function that allows each user to register the words they need individually.
- Dictionary registration work is costly and costly for individual users. For this reason, a method of sharing dictionary data registered by individual users among a plurality of users has been proposed for the purpose of reducing the cost of dictionary registration work.
- Non-Patent Document 1 describes a method in which the latest dictionaries and user dictionaries created by individual users are published on a dedicated web page and used by each user.
- this method unless individual users are constantly monitoring the homepage, it will not be possible to immediately know whether a dictionary useful for them has been published! /! There was a problem.
- Patent Document 1 dictionary data registered in the user dictionaries of a plurality of users is extracted by traversing the dictionary data registered by each user in their user dictionaries. It describes how to reflect the extracted dictionary data to the user dictionaries of all users as dictionary data to be shared!
- Patent Document 2 manages individual user dictionaries in association with specific technical term dictionaries.
- Dictionary data that is registered in common with multiple user dictionaries associated with the same technical term dictionary by examining the dictionary data registered by each user in their own user dictionary.
- a method of sharing the dictionary data among a plurality of users is described by extracting the dictionary data and reflecting the extracted dictionary data as dictionary data to be shared in the technical term dictionary.
- Patent Document 2 and Patent Document 3 also describe a method of notifying each user that new dictionary data has been registered in the shared dictionary and asking whether or not to use that word! Yes. These methods immediately inform the user that new dictionary data is available, and leave the final decision on whether or not to use the dictionary data to the user. It is intended to prevent mixing into individual user dictionaries.
- Non-Patent Document 1 NEC Corporation, “Translation Adapter II CROSSROAD Ver.3 HANDBO OK”, 1999, ⁇ .134-135
- Patent Document 1 Patent Publication 346488 IB
- Patent Document 2 Patent Public Goods SJP— 2003— 157257A
- the first problem in the prior art is that it is difficult for an individual user to make a determination as to whether or not each dictionary data is necessary. The reason for this is that even if the user is forced to use it at that time! /, Na! /, Or the dictionary data of words! / This is because it is difficult for the user to determine whether data is needed.
- the second problem is that, in the conventional method, the ability to inquire whether or not to use the dictionary data to be registered as needed from each user, and this hinders user behavior.
- the reason is that notifications and inquiries in the conventional method are accepted. This is because it is performed without considering the convenience of the user on the side of the network.
- An object of the present invention is to provide a natural language processing system that allows a user to easily determine whether or not individual dictionary data is necessary.
- Another object of the present invention is to provide a natural language processing system capable of notifying and inquiring about the necessity of dictionary data without impeding user behavior.
- the present invention compares a registration candidate storage unit that stores registration candidate dictionary data with input data and registration candidate dictionary data, and determines whether or not a word corresponding to the registration candidate dictionary data exists in the input data.
- a determination means for inquiring the user whether to register the registration candidate dictionary data in the dictionary when the determination means determines that a corresponding word exists, and an inquiry means for receiving a user instruction.
- a natural language processing system comprising a dictionary registration means for registering corresponding dictionary data in a dictionary in accordance with an instruction input to and a natural language processing means for performing natural language processing on the input data using dictionary data registered in the dictionary provide.
- the present invention compares a registration candidate storage unit that stores registration candidate dictionary data with the input data and the registration candidate dictionary data, and determines whether or not a word corresponding to the registration candidate dictionary data exists in the input data.
- a determination means for inquiring the user whether to register the registration candidate dictionary data in the dictionary when the determination means determines that a corresponding word exists, and an inquiry means for receiving a user instruction.
- a dictionary registration system comprising dictionary registration means for registering corresponding dictionary data in a dictionary in accordance with an instruction input to.
- the present invention compares input data with registration candidate dictionary data stored in a storage device to determine whether or not a word corresponding to the registration candidate dictionary data exists in the input data.
- the user is inquired whether to register the registration candidate dictionary data in the dictionary, accepts the user's instruction, and the corresponding dictionary data is registered in the dictionary according to the input instruction.
- We provide a natural language processing method that performs natural language processing by applying natural language processing to input data using dictionary data.
- the present invention compares input data with registration candidate dictionary data stored in a storage device to determine whether or not a word corresponding to the registration candidate dictionary data exists in the input data. If it is determined that the registration candidate dictionary data is present, the user is inquired whether to register the registration candidate dictionary data in the dictionary, accepts the user's instruction, and registers the corresponding dictionary data in the dictionary according to the input instructions.
- the present invention is a recording medium encoded with a computer readable computer program for operating a control arithmetic unit (CPU), and storing registration candidate dictionary data in the CPU
- the input data and the registration candidate dictionary data are compared to determine whether a word corresponding to the registration candidate dictionary data exists in the input data, and the determination processing corresponds to
- a process of inquiring the user whether to register the registration candidate dictionary data in the dictionary and receiving the user's instruction; and the registration candidate dictionary data according to the instruction input to the inquiry means Are registered in the dictionary, and natural language processing is performed on the input data using dictionary data registered in the dictionary.
- the present invention is a recording medium encoded with a computer-readable computer program for operating a control arithmetic unit (CPU), and storing registration candidate dictionary data in the CPU
- the input data and the registration candidate dictionary data are compared, a determination is made as to whether or not a word corresponding to the registration candidate dictionary data exists in the input data, and the determination process includes the correspondence If it is determined that there is a word to be registered, the user is asked whether to register the registration candidate dictionary data in the dictionary, the user's instruction is accepted, and the instruction input in the inquiry process is followed.
- a medium for executing processing for registering the registration candidate dictionary data in the dictionary is provided.
- FIG. 1 is a block diagram showing a configuration of a natural language processing system according to a first embodiment of the present invention.
- FIG. 2 is a flowchart showing the operation of the natural language processing system of FIG.
- FIG. 3 is a block diagram showing a configuration of a dictionary registration system according to a second exemplary embodiment of the present invention.
- FIG. 4 is a block diagram showing a configuration of a computer system according to a third embodiment of the present invention.
- FIG. 5 is a diagram showing an example of a data structure of a registration candidate storage unit.
- FIG. 6 is a diagram showing an example of a registration confirmation screen for dictionary data.
- FIG. 7 is a diagram showing an example of a registration confirmation screen for dictionary data.
- a natural language processing system 10 according to a first exemplary embodiment of the present invention will be described in detail with reference to the drawings.
- a natural language processing system 10 includes an input device 1 such as a keyboard and a microphone, a data processing device 2 that operates under program control, a storage device 3 that stores information, a display device, a printing device, and a speaker. Output device 4 and the like.
- the storage device 3 includes a dictionary storage unit 31 and a registration candidate storage unit 32.
- the dictionary storage unit 31 stores individual user dictionaries, and each dictionary stores words and corresponding language information! /.
- the linguistic information is information that the natural language processing means 21 to be described later refers to for the processing, and includes, for example, kana notation, reading, translation, part of speech, meaning information, and the like.
- the registration candidate storage unit 32 stores registration candidate dictionary data, which are candidates for dictionary data to be newly registered in the dictionary of each user!
- the dictionary data is the smallest unit of information registered in the dictionary, and consists of words and corresponding language information.
- the data processing device 2 includes a natural language processing means 21, a determination means 22, an inquiry means 23, and a dictionary registration means 24.
- the natural language processing means 21 accepts input from the user and performs natural language processing using the user's dictionary stored in the dictionary storage unit 31 for the input data! , Output the processing result.
- the natural language processing performed by the natural language processing means 21 is, for example, machine translation processing or speech synthesis processing.
- the machine translation process is a process of converting the input character string of the first language into a character string of the second language
- the speech synthesis process is a process of converting the input character string into a speech signal.
- the natural language processing means 21 outputs the input data to the determination means 22 before performing natural language processing on the input data.
- the natural language processing means 21 is provided with an unknown word detection function and the input data is not input to the determination means 22 only when an unknown word is found in the input data rather than always outputting the input data to the determination means 22. May be output and the determination means 22 may operate.
- the determination unit 22 compares the input data input from the natural language processing unit 21 with the words in the dictionary data stored in the registration candidate storage unit 32, and stores the input data in the registration candidate storage unit 32. Check whether the dictionary data contains the corresponding word (hereinafter referred to as “corresponding word”). When a word corresponding to the input data is found here, registration candidate dictionary data (hereinafter referred to as “corresponding dictionary data”) corresponding to the corresponding word is output to the inquiry means 23. If the determining means 22 operates only when an unknown word is found in the input data, when the determining means 22 further checks whether the word corresponding to the input data is included. Alternatively, it may be determined that the corresponding word is included only when there is an overlap between the part of the input data including the unknown word and the word of the dictionary data.
- the inquiry unit 23 displays the corresponding dictionary data input from the determination unit 22 on the output device 4 and asks the user whether or not to register it in the dictionary.
- the inquiry means 23 outputs the corresponding dictionary data to the dictionary registration means 24.
- the control is returned to the natural language processing means 21.
- the corresponding dictionary data is recorded in the registration candidate storage unit 32 together with information indicating that it is not registered, and then registered. Even if the same dictionary data is registered in the candidate storage unit 32, it may be processed as being out of registration.
- the dictionary registration unit 24 registers the corresponding dictionary data input from the inquiry unit 23 in the dictionary, and returns control to the natural language processing unit 21.
- the natural language processing means 21 receives an input from the input device 1, it outputs the input data input before executing the natural language processing to the determination means 22.
- the judging means 22 takes out the registered candidate dictionary data from the registration candidate storage section 32 (step A1 in FIG. 2).
- step A2 the ability to extract the registration candidate dictionary data from the registration candidate storage unit 32 is checked, and if it is extracted, the dictionary data is not subject to registration, and candidates for dictionary data to be registered in the dictionary are determined.
- Check for existence step A2). If there is no candidate dictionary data to be registered here! /, The natural language processing means 21 performs natural language processing on the input data and outputs the result to the output device 4 (determination in step A2). No, step A10).
- the determination means 22 compares the input data with the respective dictionary data (steps A3 and A4). If the input data contains a word corresponding to! /, NA! /, The natural language processing means 21 performs natural language processing on the input data and outputs the result to the output device 4 (determination in step A4). No, step A10).
- the inquiry means 23 displays the dictionary data (corresponding dictionary data) on the output device 4, and inquires the user whether or not to register it in the dictionary (step A5). Thereafter, an input of a response to the inquiry from the input device 1 is accepted, and it is checked whether or not there is correspondence dictionary data that has been input as not being registered (step A6). If there is corresponding dictionary data that is input as not registered, the inquiry means 23 records the dictionary data in the registered candidate storage unit 32 together with information indicating that it is not registered (step A7).
- the inquiry means 23 checks whether or not the corresponding dictionary data entered is present after registration (step A8). . If the corresponding dictionary data entered after registration exists, The book registration means 23 registers the dictionary data in the dictionary (step A9).
- the natural language processing means 21 After that, or when there is no corresponding dictionary data input when registered in step A8, the natural language processing means 21 performs natural language processing on the input data using the dictionary stored in the dictionary storage unit 31, The result is output to the output device 4 (step A10).
- the user can determine the usefulness in accordance with the context of the input data that he / she inputs only for dictionary data that is expected to be used, and can easily determine whether it is necessary.
- the user can determine whether the dictionary data is necessary or not in a series of operations that perform natural language processing on the input data entered by the user. Will be able to.
- the dictionary registration system 11 is composed of parts of the natural language processing system 10 that are necessary for registering dictionary data in the dictionary. Therefore, the parts common to the natural language processing system 10 are the same as those in FIG. The description is omitted with the symbol.
- the dictionary registration system 11 includes an unknown word search means 25 corresponding to the unknown word detection function of the natural language processing means 21 instead of the natural language processing means 21 of FIG.
- the input data input from is not a data subject to natural language processing but data such as sentences prepared in advance for dictionary registration processing.
- the output device 4 is used only for inputting an inquiry about whether or not the corresponding dictionary data can be registered and a response to the inquiry.
- the computer system 12 is similar to the first embodiment of the present invention. , Input device 1, data processing device 6, storage device 3, and output device 4.
- the natural language processing program 5 is read by the data processing device 6 to control the operation of the data processing device 6, and a dictionary storage unit 31 and a registration candidate storage unit 32 are generated in the storage device 3.
- the data processing device 6 executes the same processing as the processing by the data processing device 2 in the first embodiment under the control of the search program 5 for natural language processing.
- the dictionary registration processing program 7 may be read into the data processing device 6 instead of the natural language registration processing program 5, and the same processing as that performed by the data processing device 2 in the second embodiment may be executed! /.
- the present embodiment includes a keyboard as the input device 1, a personal computer as the data processing device 2, a magnetic disk recording device as the data storage device 3, and a display as the output device 4.
- the personal computer includes natural language processing means 21, determination means 22, and inquiry means.
- the dictionary data shown in FIG. 5 is stored in the registration candidate storage unit 32 as registration candidates.
- the contents of dictionary data as registration candidates are shown in a table format, and each line of the table represents one dictionary data.
- the first line represents dictionary data with Japanese language ability S “gene”, English “ gene ”, and part of speech “noun”.
- the fourth column stores information indicating whether the dictionary data is not registered.
- “” indicates that the inspection to determine if it is not subject to registration, which will be described later, has been completed! /, !, and “not applicable” refers to dictionary data that has been designated as not subject to inspection in the past. It represents that.
- Fig. 5 the contents of dictionary data as registration candidates are shown in a table format, and each line of the table represents one dictionary data.
- the first line represents dictionary data with Japanese language ability S “gene”, English “ gene ”, and part of speech “noun”.
- the fourth column stores information indicating whether the dictionary data is not registered.
- “” indicates that the inspection to determine if it is not
- the central processing unit extracts dictionary data as registration candidates from the registration candidate storage unit 32 and inspects the contents before performing machine translation processing on the input data. For example, by extracting dictionary data from the registration candidate storage unit 32 shown in FIG. 5 and examining the non-registration fields, the “gene diagnosis” dictionary data is out of registration, and the “gene” and “transposon” dictionaries Confirm that the data is an untested registration candidate.
- the central processing unit compares the input data input as the target of natural language processing with the dictionary data as registration candidates extracted from the registration candidate storage unit 32, and records it as "transposon” in the dictionary data. Detects that the character string (corresponding word) is included in the input data. Therefore, the central processing unit displays “transposon” dictionary data (corresponding dictionary data) on the output device 4 and asks the user whether or not to register it.
- FIG. 6 shows an example of an inquiry screen for the user. If the user checks radio button 51 of “Register” in “Transposon” and presses “Execute” button 54, the central processing unit registers the dictionary data of “Transposon” in the dictionary, and registers the candidate storage unit. Delete “Transposon” dictionary data in 32. Also, if the user checks the radio button 52 for “Do not register! /,” And presses the “execute” button 54, the central processing unit will not be registered in the dictionary data of “transposon” in the registration candidate storage unit 32. Indicate information that indicates “Not applicable” in the field. If the user selects the “Hold” radio button 53 or the “Cancel” button 55, the dictionary registration and the registration candidate storage unit 32 are not updated.
- Temporal use a radio button called “Temporary use” is added, and the displayed dictionary data is registered in the dictionary and registered in a temporary dictionary rather than permanently used, and only for natural language processing of the input data. You may take the structure which operate
- FIG. 6 a force indicating an example in which a radio button is used for a user interface for designating whether to register or not may be used as shown in FIG.
- the dictionary data with the check box 56 checked is processed in the same way as when the radio button 51 of "Register” is selected in Fig. Therefore, the same processing as when the “not register” radio button 52 is selected in FIG.
- the central processing unit performs a translation process on the input processing target, and outputs the result to the output device 4.
- the input data is compared with dictionary data.
- natural language processing is applied to the input data to detect that "transposon" or part of it is an unknown word.
- this is displayed on the output device 4 to inquire the user whether to register. It may take a form.
- the notification related to the dictionary data only when the word corresponding to the dictionary data to be registered is included in the input data input by the user to the natural language processing means.
- the user In order to make an inquiry, the user only has to determine the usefulness of the dictionary data that is expected to be used in accordance with the context of the input data that he / she entered.
- the user can determine whether or not the dictionary data is necessary in a series of operations that perform natural language processing on the input data that he / she entered, so that the user's behavior is not hindered. Can notify and inquire about the necessity of dictionary data.
- the natural language processing means has an unknown word detection function for determining whether or not the input data contains an unknown word! /, And the natural language processing means is the input data.
- the determination means may be activated when an unknown word contained in is detected.
- the determination means when the natural language processing means detects an unknown word included in the input data, the determination means includes a portion including the unknown word of the input data, registration candidate dictionary data, It is also possible to determine whether there is a word corresponding to the part containing the unknown word.
- unknown words exist in the input data. Only in the case of a fatal case, the user is notified and inquires about the corresponding word in the input data, so the user can respond to fewer inquiries and maximize the effectiveness of dictionary registration for unknown words. Obtainable.
- the natural language processing means may perform the natural language processing on the input data using only the dictionary data registered in the dictionary. .
- the natural language processing means is a machine translation means for converting a character string of the first language into a character string of the second language.
- it may be a speech synthesizer that converts an input character string into a speech signal.
- the dictionary registration system of the embodiment since the input data input by the user includes a word corresponding to the dictionary data to be registered, the user is notified and inquired about the dictionary data. Only for dictionary data that is expected to be used, the usefulness should be judged according to the context of the input data that you entered. Therefore, it is possible for the user to easily determine whether or not each dictionary data is necessary.
- the input data includes unknown word detecting means for determining whether or not an unknown word is included! /, And the unknown word detecting means is included in the input data.
- the determination means may be activated when an unknown word is detected.
- the determination unit determines the portion including the unknown word of the input data and the registration candidate dictionary data. You can also compare and determine if there is a word corresponding to the part containing the unknown word! /.
- the dictionary data is used in the machine translation process for converting the input first language character string into the second language character string.
- dictionary data used in speech synthesis processing for converting an input character string into a speech signal is used in the dictionary registration system of the embodiment.
- the dictionary is used in the machine translation process for converting the dictionary data into the second language character string from the input first language character string.
- Voice recognition that is data or dictionary data used in kana-kanji conversion processing that converts the input kana character string into a kanji-kana mixed character string, or converts the input voice signal into a character string
- the judgment means compares the input data with the registration candidate dictionary data, the input data is compared with the character string stored as a conversion result in the registration candidate dictionary data. It may be determined whether a corresponding word exists.
- the notification and inquiry regarding the dictionary data are performed only when the word corresponding to the dictionary data to be registered is included in the input data input by the user.
- the user only has to judge the usefulness of the dictionary data expected to be used in accordance with the context of the input data he / she entered. Therefore, it is possible for the user to easily determine whether or not each dictionary data is necessary.
- the user can determine whether or not the dictionary data is necessary in a series of operations that perform natural language processing on the input data that he / she entered, so that the user's behavior is not hindered. Can notify and inquire about the necessity of dictionary data.
- the registration candidate dictionary data may be compared to determine whether a word corresponding to the input data exists.
- the part including the unknown word in the input data is compared with the registration candidate dictionary data, and the unknown word is determined. It may be possible to determine whether or not a word corresponding to the included portion exists.
- an unknown word exists in the input data. Only in a fatal case, the user is notified and inquired about the corresponding word in the input data, so the user can obtain the maximum effect of dictionary registration for unknown words by responding to fewer inquiries. it can.
- the natural language processing is machine translation processing for converting a character string in the first language into a character string in the second language, or input.
- This is a speech synthesis process that converts the converted character string into a speech signal.
- the dictionary registration method of the embodiment in order to notify and inquire about the dictionary data only when the word corresponding to the dictionary data to be registered is included in the input data input by the user, The user only has to judge the usefulness of the dictionary data that is expected to be used according to the context of the input data that he / she entered.
- the input data and registration are determined when it is determined whether or not the input data includes an unknown word, and it is determined that the input data includes an unknown word. Compare the candidate dictionary data and determine whether there is a word corresponding to the input data.
- the dictionary registration method of the embodiment when it is determined that the input data includes an unknown word, the part including the unknown word in the input data is compared with the registration candidate dictionary data, and the unknown word is determined. You may make it determine whether the word corresponding to the part to include exists. According to the dictionary registration method of the above-described embodiment, only when a fatal case where an unknown word is present in the input data is fatal, a notification or inquiry about the corresponding word existing in the input data is performed, so the user can The effect of dictionary registration for unknown words can be maximized by responding to a small number of inquiries.
- the dictionary data is dictionary data used in a machine translation process for converting an input character string of a first language into a character string of a second language.
- a dictionary used in speech synthesis processing that converts input character strings into speech signals It may be written data.
- the notification related to the dictionary data only when the word corresponding to the dictionary data to be registered is included in the input data input by the user, Since the computer is operated so as to make an inquiry, the user only has to determine the usefulness of the dictionary data that he / she expects to use according to the context of the input data he / she entered.
- the user can determine whether or not the dictionary data is necessary in a series of operations that perform natural language processing on the input data that he / she entered, so that the user's behavior is not hindered. Can notify and inquire about the necessity of dictionary data.
- the natural language processing function has a function of determining whether or not an unknown word is included in the input data. If it is determined that an unknown word is included, the determination process may be executed by a computer.
- the determination process includes the unknown word of the input data.
- registration candidate dictionary data may be compared to determine whether or not there is a word corresponding to a portion including an unknown word.
- the input data Only in the case of a fatal case where an unknown word exists in the data, the user is notified and inquired about the corresponding word existing in the input data, so the user can register the dictionary for the unknown word by responding to fewer inquiries. You can get the maximum effect.
- the natural language processing function may perform natural language processing on input data using only dictionary data registered in the dictionary.
- the natural language processing function is a machine translation function that converts a character string of the first language into a character string of the second language, Or as a voice synthesis function that converts the input character string into a voice signal! /.
- the dictionary registration program of the embodiment only when the word corresponding to the dictionary data to be registered is included in the input data input by the user, the notification and inquiry regarding the dictionary data are made.
- the user In order to make the computer operate like this, the user should only determine the usefulness of the dictionary based on the input data entered for the dictionary data that is expected to be used.
- the medium in which the dictionary registration program of the embodiment is recorded! / causes the computer to execute an unknown word detection process for determining whether or not an unknown word is included in the input data, thereby detecting the unknown word.
- the determination process may be executed by a computer.
- the determination process determines that the input data is unknown. Compare the part containing the word with the registration candidate dictionary data to determine whether there is a word corresponding to the part containing the unknown word! /.
- the dictionary data may be used in the dictionary, or may be dictionary data used in a voice synthesis function that converts an input character string into a voice signal.
- a machine translation function for converting dictionary data into a second language character string by converting the dictionary data into a medium recording the dictionary registration program of the embodiment! Is dictionary data used by the Kana / Kanji conversion function that converts the input kana character string to a kanji-kana mixed character string, or the input audio signal as a character string
- the input data and the registration candidate dictionary data are compared. Make sure that the corresponding word exists semi-IJ.
- the first effect achieved in the preferred embodiment of the present invention is that the user can easily determine whether or not each dictionary data is necessary.
- the reason is that notification and inquiry regarding the dictionary data are performed only when words corresponding to the dictionary data to be registered are included in the input data input to the natural language processing means by the user. This is because the user only has to judge the usefulness of the dictionary data that he / she expects to use according to the context of the input data he / she entered.
- the second effect is that it is possible to notify and inquire about the necessity of dictionary data without obstructing user behavior.
- the reason is to notify and inquire only when the dictionary data to be registered is included in the input data entered by the user into the natural language processing means. This is because the user can determine whether the dictionary data is necessary or not in a series of operations for performing natural language processing on the input data input by the user.
- the present invention provides a computer with a mechanical translation device that converts an input character string of a first language into a character string of a second language, and a speech synthesizer that converts the input character string into a speech signal. It can be applied to uses such as a program for realizing. In the embodiment, the example is shown in which the user starts the operation by inputting the data to be processed by the natural language processing. In addition, a program for implementing a dictionary creation support device that supports creation of a dictionary used in natural language processing on a computer by adopting a configuration that starts operation by inputting data that is not directly subject to natural language processing. It can also be applied to applications.
- an operation is started by inputting a document to be obtained as a result of natural language processing in the future, and this input data is compared with a character string stored as a result of natural language processing in dictionary data.
- a configuration that determines whether or not the dictionary data included in the input data exists for example, creation of a dictionary used in machine translation processing, kana-kanji conversion processing, voice recognition processing, and! /, Natural language processing It can also be applied to applications such as a program for realizing a dictionary creation support device on a computer.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Machine Translation (AREA)
- Document Processing Apparatus (AREA)
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2008533198A JP5239863B2 (ja) | 2006-09-07 | 2007-09-06 | 自然言語処理システムおよび辞書登録システム |
CN200780033343.6A CN101512518B (zh) | 2006-09-07 | 2007-09-06 | 自然语言处理系统和词典登录系统 |
US12/310,773 US9575953B2 (en) | 2006-09-07 | 2007-09-06 | Natural-language processing system and dictionary registration system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2006-242519 | 2006-09-07 | ||
JP2006242519 | 2006-09-07 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2008029881A1 true WO2008029881A1 (fr) | 2008-03-13 |
Family
ID=39157306
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2007/067416 WO2008029881A1 (fr) | 2006-09-07 | 2007-09-06 | Système de traitement du langage naturel et système d'enregistrement de dictionnaire |
Country Status (4)
Country | Link |
---|---|
US (1) | US9575953B2 (ja) |
JP (1) | JP5239863B2 (ja) |
CN (1) | CN101512518B (ja) |
WO (1) | WO2008029881A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20140077788A (ko) * | 2012-12-14 | 2014-06-24 | 한국전자통신연구원 | 음성인식 시스템에서 유사도를 기반으로 한 비인식 대상 단어 생성 방법 |
KR20200036419A (ko) * | 2018-09-28 | 2020-04-07 | 우석대학교 산학협력단 | 음성을 이용한 키워드 추출 방법 및 서버 |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9575953B2 (en) * | 2006-09-07 | 2017-02-21 | Nec Corporation | Natural-language processing system and dictionary registration system |
JP2009099056A (ja) * | 2007-10-18 | 2009-05-07 | Panasonic Corp | 文字列受信装置、文字列転送装置、文字列送受信システム、コンテンツ受信端末向けシステムlsi、名簿共有システム、名簿共有方法およびコンテンツ推薦方法 |
JP5335227B2 (ja) * | 2007-12-10 | 2013-11-06 | 京セラ株式会社 | 情報端末装置 |
JP2015060095A (ja) * | 2013-09-19 | 2015-03-30 | 株式会社東芝 | 音声翻訳装置、音声翻訳方法およびプログラム |
JP6447068B2 (ja) * | 2014-12-05 | 2019-01-09 | カシオ計算機株式会社 | 情報処理装置、情報表示方法及びプログラム |
KR102117082B1 (ko) | 2014-12-29 | 2020-05-29 | 삼성전자주식회사 | 음성 인식 방법 및 음성 인식 장치 |
US11250872B2 (en) * | 2019-12-14 | 2022-02-15 | International Business Machines Corporation | Using closed captions as parallel training data for customization of closed captioning systems |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10260984A (ja) * | 1997-03-19 | 1998-09-29 | Toshiba Corp | 辞書管理方法および辞書管理装置および辞書利用システム |
Family Cites Families (50)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPS5840684A (ja) * | 1981-09-04 | 1983-03-09 | Hitachi Ltd | 自然言語間の自動翻訳方式 |
GB2199170A (en) * | 1986-11-28 | 1988-06-29 | Sharp Kk | Translation apparatus |
JPH01234975A (ja) * | 1988-03-11 | 1989-09-20 | Internatl Business Mach Corp <Ibm> | 日本語文章分割装置 |
JP2836159B2 (ja) * | 1990-01-30 | 1998-12-14 | 株式会社日立製作所 | 同時通訳向き音声認識システムおよびその音声認識方法 |
JPH03268062A (ja) * | 1990-03-19 | 1991-11-28 | Fujitsu Ltd | 機械翻訳電子メール装置における私用単語の登録装置 |
JPH0594436A (ja) * | 1990-10-10 | 1993-04-16 | Fuji Xerox Co Ltd | 文書処理装置 |
US5497319A (en) * | 1990-12-31 | 1996-03-05 | Trans-Link International Corp. | Machine translation and telecommunications system |
US5742834A (en) * | 1992-06-24 | 1998-04-21 | Canon Kabushiki Kaisha | Document processing apparatus using a synonym dictionary |
US5867812A (en) * | 1992-08-14 | 1999-02-02 | Fujitsu Limited | Registration apparatus for compound-word dictionary |
JPH06314274A (ja) * | 1993-04-28 | 1994-11-08 | Toshiba Corp | 文書作成装置及び文書情報入力方法 |
JP3408291B2 (ja) * | 1993-09-20 | 2003-05-19 | 株式会社東芝 | 辞書作成支援装置 |
JPH07175813A (ja) * | 1993-10-27 | 1995-07-14 | Ricoh Co Ltd | 複合通信処理装置 |
JPH07182465A (ja) * | 1993-12-22 | 1995-07-21 | Hitachi Ltd | 文字認識方法 |
TW274135B (ja) * | 1994-09-14 | 1996-04-11 | Hitachi Seisakusyo Kk | |
KR970701886A (ko) * | 1994-11-21 | 1997-04-12 | 다떼이시 요시오 | 복수의 클라이언트 장치에 공유되는 데이터베이스 시스템, 데이터 갱신방법 및 문자 처리장치에의 응용(Data base system shared by a plurality of client apparatuses, data updating method and application to character processor) |
JPH08167006A (ja) * | 1994-12-13 | 1996-06-25 | Canon Inc | 自然言語処理装置及びその方法 |
US5649221A (en) * | 1995-09-14 | 1997-07-15 | Crawford; H. Vance | Reverse electronic dictionary using synonyms to expand search capabilities |
JPH09114828A (ja) * | 1995-10-20 | 1997-05-02 | Sanyo Electric Co Ltd | 変換候補の優先順位情報の更新方法 |
JPH09161014A (ja) * | 1995-12-13 | 1997-06-20 | Oki Electric Ind Co Ltd | 文字認識装置 |
JP3254642B2 (ja) * | 1996-01-11 | 2002-02-12 | 株式会社日立製作所 | 索引の表示方法 |
US6085162A (en) * | 1996-10-18 | 2000-07-04 | Gedanken Corporation | Translation system and method in which words are translated by a specialized dictionary and then a general dictionary |
JPH10198680A (ja) * | 1997-01-07 | 1998-07-31 | Hitachi Ltd | 分散辞書管理方法及びそれを用いた機械翻訳方法 |
US5924068A (en) * | 1997-02-04 | 1999-07-13 | Matsushita Electric Industrial Co. Ltd. | Electronic news reception apparatus that selectively retains sections and searches by keyword or index for text to speech conversion |
JP3464881B2 (ja) | 1997-03-25 | 2003-11-10 | 株式会社東芝 | 辞書構築装置および方法 |
US6081774A (en) * | 1997-08-22 | 2000-06-27 | Novell, Inc. | Natural language information retrieval system and method |
US6507678B2 (en) * | 1998-06-19 | 2003-01-14 | Fujitsu Limited | Apparatus and method for retrieving character string based on classification of character |
US6401060B1 (en) * | 1998-06-25 | 2002-06-04 | Microsoft Corporation | Method for typographical detection and replacement in Japanese text |
US7254531B2 (en) * | 2000-09-05 | 2007-08-07 | Nir Einat H | In-context analysis and automatic translation |
JP2002108858A (ja) * | 2000-09-20 | 2002-04-12 | Internatl Business Mach Corp <Ibm> | 機械翻訳方法、機械翻訳装置および記録媒体 |
US6973427B2 (en) * | 2000-12-26 | 2005-12-06 | Microsoft Corporation | Method for adding phonetic descriptions to a speech recognition lexicon |
US7254530B2 (en) * | 2001-09-26 | 2007-08-07 | The Trustees Of Columbia University In The City Of New York | System and method of generating dictionary entries |
NO316480B1 (no) * | 2001-11-15 | 2004-01-26 | Forinnova As | Fremgangsmåte og system for tekstuell granskning og oppdagelse |
JP3838904B2 (ja) | 2001-11-22 | 2006-10-25 | 沖電気工業株式会社 | 辞書装置及び自然言語処理システム |
JP4650920B2 (ja) * | 2002-04-16 | 2011-03-16 | 富士通株式会社 | 情報処理装置及び情報処理プログラム |
KR100530154B1 (ko) * | 2002-06-07 | 2005-11-21 | 인터내셔널 비지네스 머신즈 코포레이션 | 변환방식 기계번역시스템에서 사용되는 변환사전을생성하는 방법 및 장치 |
US7016895B2 (en) * | 2002-07-05 | 2006-03-21 | Word Data Corp. | Text-classification system and method |
US7548863B2 (en) * | 2002-08-06 | 2009-06-16 | Apple Inc. | Adaptive context sensitive analysis |
JP4355138B2 (ja) * | 2002-12-13 | 2009-10-28 | インターナショナル・ビジネス・マシーンズ・コーポレーション | 翻訳サーバ、コラボレーションサーバ及びプログラム |
US7437296B2 (en) * | 2003-03-13 | 2008-10-14 | Matsushita Electric Industrial Co., Ltd. | Speech recognition dictionary creation apparatus and information search apparatus |
US20070174041A1 (en) * | 2003-05-01 | 2007-07-26 | Ryan Yeske | Method and system for concept generation and management |
CN100429648C (zh) * | 2003-05-28 | 2008-10-29 | 洛昆多股份公司 | 一种文本自动分块的方法、分块器和文本到语言合成系统 |
JP4652737B2 (ja) * | 2004-07-14 | 2011-03-16 | インターナショナル・ビジネス・マシーンズ・コーポレーション | 単語境界確率推定装置及び方法、確率的言語モデル構築装置及び方法、仮名漢字変換装置及び方法、並びに、未知語モデルの構築方法、 |
CN100568230C (zh) * | 2004-07-30 | 2009-12-09 | 国际商业机器公司 | 基于超文本的多语言网络信息搜索方法和系统 |
JP4570509B2 (ja) * | 2005-04-22 | 2010-10-27 | 富士通株式会社 | 読み生成装置、読み生成方法及びコンピュータプログラム |
JP2007033901A (ja) * | 2005-07-27 | 2007-02-08 | Nec Corp | 音声認識システム、音声認識方法、および音声認識用プログラム |
US7620549B2 (en) * | 2005-08-10 | 2009-11-17 | Voicebox Technologies, Inc. | System and method of supporting adaptive misrecognition in conversational speech |
JP2007058509A (ja) * | 2005-08-24 | 2007-03-08 | Toshiba Corp | 言語処理システム |
JP2007206975A (ja) * | 2006-02-01 | 2007-08-16 | Toshiba Corp | 言語情報変換装置及びその方法 |
US7899664B2 (en) * | 2006-05-22 | 2011-03-01 | Sharp Kabushiki Kaisha | Information processing apparatus, computer, information processing system, information processing method, and program for receiving a character string and returning conversion candidates |
US9575953B2 (en) * | 2006-09-07 | 2017-02-21 | Nec Corporation | Natural-language processing system and dictionary registration system |
-
2007
- 2007-09-06 US US12/310,773 patent/US9575953B2/en active Active
- 2007-09-06 CN CN200780033343.6A patent/CN101512518B/zh active Active
- 2007-09-06 JP JP2008533198A patent/JP5239863B2/ja active Active
- 2007-09-06 WO PCT/JP2007/067416 patent/WO2008029881A1/ja active Search and Examination
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH10260984A (ja) * | 1997-03-19 | 1998-09-29 | Toshiba Corp | 辞書管理方法および辞書管理装置および辞書利用システム |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20140077788A (ko) * | 2012-12-14 | 2014-06-24 | 한국전자통신연구원 | 음성인식 시스템에서 유사도를 기반으로 한 비인식 대상 단어 생성 방법 |
KR101711941B1 (ko) * | 2012-12-14 | 2017-03-03 | 한국전자통신연구원 | 음성인식 시스템에서 유사도를 기반으로 한 비인식 대상 단어 생성 방법 |
KR20200036419A (ko) * | 2018-09-28 | 2020-04-07 | 우석대학교 산학협력단 | 음성을 이용한 키워드 추출 방법 및 서버 |
KR102166446B1 (ko) | 2018-09-28 | 2020-10-15 | 우석대학교 산학협력단 | 음성을 이용한 키워드 추출 방법 및 서버 |
Also Published As
Publication number | Publication date |
---|---|
US20090281786A1 (en) | 2009-11-12 |
CN101512518B (zh) | 2015-06-24 |
US9575953B2 (en) | 2017-02-21 |
JP5239863B2 (ja) | 2013-07-17 |
JPWO2008029881A1 (ja) | 2010-01-21 |
CN101512518A (zh) | 2009-08-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2008029881A1 (fr) | Système de traitement du langage naturel et système d'enregistrement de dictionnaire | |
JP4416643B2 (ja) | マルチモーダル入力方法 | |
JP2014010623A (ja) | 音声翻訳装置、方法およびプログラム | |
WO2008113717A1 (en) | Disambiguating text that is to be converted to speech using configurable lexeme based rules | |
JPH11338858A (ja) | 入力予測装置、入力予測方法及び入力予測プログラムを記録した記録媒体 | |
JP2007219190A (ja) | 音声認識装置と認識方法及びそのプログラム | |
JPH05324702A (ja) | 情報処理装置 | |
JPWO2007066433A1 (ja) | 音声認識装置 | |
JPH08314910A (ja) | 異種コード文字列転記装置および電子辞書 | |
JP2006065651A (ja) | 商標称呼検索プログラム、商標称呼検索装置及び商標称呼検索方法 | |
JP4643183B2 (ja) | 翻訳装置および翻訳プログラム | |
JP2008225774A (ja) | 情報処理装置及びこの装置の階層メニュー表示方法並びに階層メニュー表示プログラム | |
JP4300056B2 (ja) | 概念表現生成方法、プログラム、記憶媒体及び概念表現生成装置 | |
JP2000285112A (ja) | 予測入力装置及び予測入力方法並びに記録媒体 | |
JP2005208825A (ja) | データ検索システム、データ検索方法及びプログラム | |
JP2015072701A (ja) | 翻訳装置、方法およびプログラム | |
JP2006155529A (ja) | 辞書登録装置、辞書登録方法および辞書登録プログラム | |
JP2010055506A (ja) | 音声翻訳装置および方法 | |
Troxel | Exploring the Isaiah Scrolls and Their Textual Variants by Donald W. Parry | |
JP3410163B2 (ja) | 言語処理装置 | |
JP5363522B2 (ja) | 日本語の単字入力を支援する機能を提供する方法及びシステム | |
JP2006155262A (ja) | 入力支援システム、入力支援方法及びプログラム | |
JP2004334728A (ja) | 漸進的文解釈支援装置、及び漸進的文解釈支援プログラム | |
JPH083795B2 (ja) | 質問推定装置 | |
JPH04116763A (ja) | 漢字読み変換方式 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 200780033343.6 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 07806857 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2008533198 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12310773 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 07806857 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) |