JP2003060790A - Speech interactive translation service system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a speech interactive translation service system for writing an intermediate language processing by an XML and the like and supplying service by natural interaction to a user who has only a fixed telephone set or a portable telephone set in language that the user desires. SOLUTION: One or more speech response servers connected to a public line network, one or more translation servers and one or more service supply servers are interconnected through a network. The translation server translates service held in the service supply server, and the speech response server provides the translated result by means of the speech interaction.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice dialogue translation service system, and more particularly to a system for providing a voice dialogue translation service by voice recognition, translation and voice synthesis to a user from a telephone.

[0002]

2. Description of the Related Art As a conventional translation service system, for example, a "multinational service providing system" disclosed in Japanese Patent Laid-Open No. 9-81569 is known. In this conventional system, a conversion rule management server that manages rules for converting information representation formats via an intermediate representation format is connected to a communication network to which a service server and service clients are connected, and the service server and service clients are connected. Has a means for converting the information representation format using conversion rules, and is configured to convert a word in a specific language into an intermediate expression when transmitting, and convert the intermediate expression into a word in a specific language when receiving. It had been. Here, by providing the client with the voice synthesis means and the voice recognition means,
Service by voice dialogue was also possible.

[0003]

However, the above system is premised on an apparatus having a key input and a screen output such as a personal computer as a service server and a client, and a user who owns only a fixed telephone or a mobile telephone or There is a problem that the service provider cannot use it.

Further, there is a problem in that the intermediate representation is only provided as a voice dialogue by the voice synthesizing means and the voice recognizing means, and it cannot always be a natural dialogue that is easy to use.

Therefore, the present invention solves the above problems of the conventional system and can be used by a user or a service provider who owns only a fixed telephone or a mobile telephone, and in a language desired by the user. It is an object of the present invention to provide a service by a user-friendly natural dialogue and to provide a voice dialogue translation service system capable of coping with it.

[0006]

A speech dialogue translation service system of the present invention for solving the above-mentioned object includes one or a plurality of voice response servers connected to a public line network.
One or a plurality of translation servers and one or a plurality of service providing servers are connected by a network, the services held in the service providing server are translated by the translation server, and the result is transmitted by the voice response server. It is characterized by being provided by voice dialogue.

The voice response server may be provided with a dialogue sequence generation means for analyzing an intermediate language and generating a dialogue sequence of a specific language, for example, XML or VoiceXML.

The dialogue sequence generation means may be configured to hold a dialogue sequence generation rule for each language.

The voice response server is provided with an input / output unit for outputting a voice based on a dialogue sequence and extracting an element of an intermediate language from a voice of a specific language uttered by the user or a push input of the user. Good. Further, a service management means for managing which service providing server the intermediate language corresponding to the service exists may be provided.

A service holding means for holding an intermediate language corresponding to a service to be provided may be provided in the service providing server. The intermediate language may be composed of a language-common portion and a language-dependent portion.

When the voice response server receives a call for a service providing request including designation of a service and a service providing language from a user, the voice response server accesses the corresponding service providing server via a network to obtain a service name and a service providing language. When requesting an intermediate language by specifying, and receiving the intermediate language from the service providing server, necessary information is extracted from the intermediate language, and a dialogue sequence is generated based on the information by using the dialogue sequence creating means. , It may be configured to interact with the user according to the dialogue sequence.

Further, the voice response server may be configured to determine a service providing server holding an intermediate language corresponding to a service designated by the service managing means and access the corresponding service providing server. OK, when the user interacts with the user according to the dialog sequence, the user's utterance content is acquired, the information acquired by the dialog is substituted as the value of the language-dependent part of the intermediate language, and the value is substituted. The intermediate language may be transmitted to the service providing server.

When there is a request for an intermediate language from the voice response server in the previous term, the service providing server holds the intermediate language corresponding to the requested service name and the service providing language, and if it holds the intermediate language, the service providing server sends it to the voice response server. And the setting language of the intermediate language corresponding to the requested service name is different from the requested service providing language, the service provision requested from the intermediate language corresponding to the requested service name is performed using the translation server. The intermediate language corresponding to the language may be created and then sent to the voice response server.

Further, when the service providing server creates the intermediate language corresponding to the requested service providing language from the intermediate language corresponding to the requested service name, the created intermediate language is a value in the original setting language. And the translated value may be distinguished and described. Further, when the intermediate language is received from the previous voice response server, it may be configured to be provided to the service provider. .

Further, in the service providing server, when providing the information of the user to the service provider, the part depending on the language of the intermediate language is composed of the value of the original setting language and the value of the translated setting language. In this case, the part corresponding to the original setting language may be selected and provided to the service provider.

Further, when the service providing server sends the intermediate language to the voice response server, the service providing server also sends a dialogue sequence generation rule corresponding to the designated language, and the dialogue sequence creating means sends the dialogue sequence creating means from the service providing server. The dialogue sequence generation rule may be received, and the received dialogue sequence generation rule may be used when the dialogue sequence is generated.

The voice response server may have a function of the service providing server.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of a speech dialogue translation service system according to the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing a first embodiment of a speech dialogue translation service system according to the present invention. A plurality of voice response servers 1, a plurality of translation servers 2 and a plurality of service providing servers 6 are connected to the network 3. Each voice response server 1
Is connected to a plurality of telephones 5 via the public network 4,
Each service providing server 6 is connected to the computer 7 via a network.

FIG. 2 shows the configuration of the voice response server 1 shown in FIG.
FIG. 3 is a diagram showing a configuration of the service providing server 6. The voice response server 1 generates a dialogue sequence of a specific language from the line control means 11 for controlling the line connection with the telephone 5 on the user side via the public line network 4 and the intermediate language exchanged between the voice response servers. Dialog sequence generation means 12
And an input / output unit 13 that outputs a voice based on a dialogue sequence and extracts an element of an intermediate language from a voice of a specific language uttered by the user or a push input of the user, and which service is an intermediate language corresponding to the service. It has a service management means 14 for managing whether it exists in the providing server and an intermediate language analysis means 15 for analyzing the intermediate language and extracting information.

Further, the dialogue sequence generation means 12 has a dialogue sequence generation rule 16 (not shown) for generating a dialogue sequence from the result of analyzing the intermediate language.

The translation server 2 has a conversion dictionary 17 (not shown) in which correspondences between words in a certain language and words in another language are described.

The service providing server 6 has an intermediate language analyzing means 18 for analyzing the intermediate language and extracting information, and a service holding means 19 for holding the intermediate language.

First of all, the user operates the telephone 5 to the public line network 4
One voice server 1 is called via, and the service providing language and the service are specified by push input or voice input. For example, "1 #" (Japanese), "4
You can specify the hotel reservation service in Japanese by either pressing "#" (Hotel reservation) or saying "Japanese" or "Hotel". For calls originating from within Japan, if nothing is specified, Japanese may be specified as the language provided.

The voice response server 1 is the service management means 1
4 is used to determine in which service providing server the intermediate language corresponding to the specified service exists. After that, the service providing server 6 is accessed via the network 3, the service name and the service providing language are designated, and the intermediate language is requested.

If the service providing server 6 holds an intermediate language corresponding to the requested service name and service providing language, it returns it to the voice response server 1. When the setting language of the intermediate language corresponding to the requested service name is different from the requested service providing language, the request is made from the intermediate language corresponding to the requested service name using the intermediate language analysis means 18 and the translation server 2. After creating an intermediate language corresponding to the provided service providing language, it is returned to the voice response server 1.

Here, the intermediate language is made up of tags common to the language, values dependent on the language, and other information, as shown in FIG. 4, for example.

In FIG. 4, elements enclosed in <> indicate tags common to languages, and elements enclosed by tags indicate their values. Tags can be nested. For example, <language> ENG </ langua
ge> means that the setting language is ENG (English),
In the lines from <city> to </ city>, the value of the city name is not set, and the options are described as a vocabulary list in the file (city.lst) shown in FIG. 5, for example, <date name = "A_da
The three lines from te ″> to </ date> are da
a_da expressed by te of arrival
The value of the variable te has not been set, <attr
The line from ibut> to </ attribute> is represented by type of room r
The value of the variable "oom_type" is not set, and it indicates that there are single and twin as options. In this example, the intermediate language is described in the XML format, but it is not limited to this and a dedicated description format may be used.

The translation of the intermediate language is performed as follows. First, the service providing server 6 that holds the intermediate language
Makes a copy of the intermediate language. The service providing server 6 analyzes the copy of the intermediate language in order from the beginning by using the intermediate language analysis unit 18, and extracts the tag, the value of the tag, and other information. If the extracted tag is a language setting tag, reset the value to the target language and add the original language setting as another tag. In the case of other tags, the extracted value is sent to the translation server 2 via the network 3 together with the set language of the intermediate language and the flag of the target language to be translated to request translation. And receive the translated value,
Add the translated value to the tag value separately from the original value. If there is a file specified as a list of vocabulary in the tag, the extension of the target language is added to the file name.

The translation server 2 uses the conversion dictionary 17 corresponding to the flags of the set language of the intermediate language and the target language sent.
Translate the value sent using (not shown) into the target language,
The translated value is returned to the service providing server 6 together with the original value.

FIG. 6 shows an example of the intermediate language after translation, and FIG. 7 shows a vocabulary list file after translation. 6 and 7, values in {} are values in the original language (ENG; English).

Upon receiving the intermediate language from the service providing server 6, the voice response server 1 analyzes the intermediate language in order from the beginning using the intermediate language analysis means 15 and extracts the tag, the tag value, and other information. For example, in FIG. 6, <l
The value for the "annuage>tag" is JPN (Japanese),
The value for the <org_language> tag is ENG
(English), the value for the <city> tag is not set, and the options are city. lst. The information that is described in the JPN file is obtained. As for the <date> tag, the information that the variable a_date, the Japanese notation “arrival date”, the English notation dateof arrival1, and the value are not set is obtained. Thereafter, a dialogue sequence is generated from the tags and tag values extracted by the dialogue sequence generation means 12 and other information.

The dialogue sequence generation means 12 generates a dialogue sequence using a dialogue sequence generation rule 16 (not shown) which describes the dialogue sequence generation rule for each tag. FIG. 8 shows an example of the dialogue sequence generation rule 16 corresponding to Japanese, and FIG. 9 shows an example of the dialogue sequence generated.

In FIG. 8, setgram is a recognition vocabulary setting, prompt is a voice output, and getvar.
Indicates voice recognition or DTMF input, submit means transmission of recognition result, and $ var and $ exp represent variables. Further, in the example of FIG. 9, the dialogue sequence is described in VoiceXML, but it is not limited to this and a dedicated description method may be used. In addition, XML
To VoiceXML is described in Information Processing Society of Japan, Research Report 2000-SLP-34 "XML-Voic".
The method described in "Development of eXML conversion tool" can be used.

After creating the dialogue sequence, the voice response server 1 uses the line control means 11 and the input / output means 13 to have a dialogue with the user according to the dialogue sequence, obtain the content of the user's utterance, and tag the intermediate language. Substitute as the value of. For example, in the dialogue sequence shown in FIG. 9, the voice response server 1 outputs the voice as "Please specify the city name", and outputs the user's utterance content as city. lst. Recognize as one of the list of city names described in JPN, and substitute the result into the value of the <city> tag of the intermediate language. At that time, the value of the original language described in {} of the word list file is also substituted. For example, the recognition result is "Atlanta"
In the case of, the value to be substituted is “Atlanta
a} ”. Similarly, the arrival date acquired by the dialogue is expressed as (yyyy) mmdd in the variable a_date <d.
ate> tag, the departure date in the format of (yyyy) mmdd, the <date> tag of the variable a_date, and the room type are substituted for the values of the <attribute> tag of the variable room_type. The intermediate language after the dialogue is as shown in FIG.

After that, the voice response server 1 sends the intermediate language in which the value is substituted to the service providing server 6. Upon receiving the intermediate language, the service providing server 6 again extracts the tag, the value of the tag, and other information by using the intermediate language analysis means 18, selects only the value and information corresponding to the original set language, and selects the computer. Provide to 7.

As described above, according to the first embodiment of the present invention, since it has the above-described operation and configuration, even for a user who owns only a fixed telephone or a mobile telephone, It is possible to provide a service by a natural dialogue in a language desired by the user.

(Second Embodiment) FIG. 11 is a block diagram showing the second embodiment of the speech dialogue translation service system according to the present invention. A plurality of voice response servers 1, a plurality of translation servers 2 and a plurality of service providing servers 6 are connected via a network 3. Each voice response server 1 is connected to a plurality of telephones 5 via a public line network 4.

FIG. 12 shows the configuration of the voice response server 1 according to the second embodiment. Voice response server 1
Is a line control means 11 for controlling the line connection with the telephone 5 on the user side via the public line network 4, and a dialogue sequence generation means for generating a dialogue sequence of a specific language from an intermediate language exchanged between the voice response servers. 12, an input / output unit 13 for outputting a voice based on a dialogue sequence and extracting an intermediate language element from a voice of a specific language uttered by the user or a user's push human power, and an intermediate language corresponding to the service. It has a service management means 14 for managing whether it exists in the service providing server, an intermediate language analysis means 15 for analyzing the intermediate language and extracting information, and a service holding means 19.

Here, the service holding means 19 is the same as the service holding means in the service providing server 6 in the first embodiment.

The user calls one voice response server 1 from the telephone 5 through the public network 4 (in order to distinguish it from other voice response servers, this is referred to as "1A") and specifies the language and service. Then, the voice response server 1A accesses the voice response server 1 (which is referred to as "1B") holding the desired service, specifies the service name and the service providing language, and requests the intermediate language. Here, the voice response server 1A may hold a desired service, that is, the voice response server 1B may be the same as the voice response server 1A.

Thereafter, the voice response server 1B operates in the same way as the service providing server 6 in the first embodiment, and the voice response server 1A operates in the voice response server 1 in the first embodiment.
The same operation as in the first embodiment is performed until the voice response server 1B receives the intermediate language in which the value is substituted.

Thereafter, the voice response server 1B analyzes the intermediate language in order from the beginning by using the intermediate language analysis means 15 and extracts the tag, the value of the tag, and other information. Then, the dialogue sequence is generated from the tag, the tag value, and other information extracted by using the dialogue sequence generation means 12.

After creating the dialogue sequence, the voice response server 1B uses the line control means 11 and the input / output means 13 to have a dialogue with the service provider according to the dialogue sequence and provide the user information.

As described above, according to the second embodiment of the present invention, the service holding means 1 is provided in the voice response server 1.
9 is provided so as to have the function of the service providing server 6 in the first embodiment as well, so that even a user who has only a fixed-line telephone or a mobile telephone can use a natural language in the user's desired language. We can provide services by dialogue,
Moreover, it can respond to it.

The preferred embodiments of the speech dialogue translation service system of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to these examples. It is obvious that a so-called person skilled in the art can come up with various changes or modifications within the scope of the technical idea described in the claims, and of course, the technical scope of the present invention is also applicable to them. Be understood to belong to.

[0047]

According to the speech dialogue translation service system of the present invention, a speech dialogue system by natural dialogue is realized by using speech recognition and speech synthesis, and intermediate language processing is described by XML or the like. You can easily increase the variation of the fixed text.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a first embodiment of a system according to the present invention.

FIG. 2 is a block diagram of a voice response server of the system according to the present invention.

FIG. 3 is a configuration diagram of a service providing server of the system according to the present invention.

FIG. 4 is an example of an intermediate language of the system according to the invention.

FIG. 5 is an example of a vocabulary list file of the system according to the present invention.

FIG. 6 is an example of the translated intermediate language of the system according to the present invention.

FIG. 7 is an example of a translated vocabulary list file of the system according to the present invention.

FIG. 8 is an example of a dialogue sequence generation rule corresponding to Japanese in the system according to the present invention.

FIG. 9 is an example of a dialogue sequence generated by the system according to the present invention.

FIG. 10 is an example of an intermediate language after dialogue in the system according to the present invention.

FIG. 11 is a configuration diagram showing a second embodiment of a system according to the present invention.

FIG. 12 is a diagram showing a configuration of a voice response server in the second exemplary embodiment of the system according to the present invention.

[Explanation of symbols]

1 Voice response server 2 Translation server 3 network 4 public network 5 telephone 6 Service providing server 7 computer 11 Line control means 12 Dialog sequence generation means 13 Input / output means 14 Service management means 15 Intermediate language analysis means 16 Dialog sequence generation rules 17 conversion dictionary 18 Intermediate language analysis means 19 Service holding means

Front page continuation (51) Int.Cl. ⁷ Identification code FI theme code (reference) H04M 3/523 G10L 3/00 551C 5K101 11/10 571V R F term (reference) 5B091 AA01 BA13 CB12 CB32 CD03 EA09 EA21 5D015 KK02 LL12 5D045 AB03 AB04 AB24 5K015 AA06 AA07 GA07 5K024 BB01 BB02 DD01 DD02 5K101 KK15 MM07 NN07 NN16

Claims (16)

[Claims] 1. A service providing server in which one or more voice response servers connected to a public line network, one or more translation servers, and one or more service providing servers are connected by a network. A voice dialogue translation service system, wherein the translation server translates the service held therein, and the voice response server provides the result by voice dialogue. 2. The voice dialogue translation service system according to claim 1, wherein the voice response server has dialogue sequence generation means for analyzing an intermediate language to generate a dialogue sequence of a specific language. 3. The voice dialogue translation service system according to claim 2, wherein the dialogue sequence generation means holds a dialogue sequence generation rule for each language. 4. The voice response server has an input / output unit that outputs a voice based on a dialogue sequence and extracts an intermediate language element from a voice of a specific language uttered by the user or a push input of the user. The spoken dialogue translation service system according to any one of claims 1 to 3. 5. The voice response server has service management means for managing which service providing server the intermediate language corresponding to the service is present in, as claimed in any one of claims 1 to 4. The spoken dialogue translation service system described. 6. The spoken dialogue translation service system according to claim 1, wherein the service providing server has a service holding unit that holds an intermediate language corresponding to the service to be provided. . 7. The spoken dialogue translation service system according to claim 1, wherein the intermediate language is composed of a language-common portion and a language-dependent portion. . 8. The voice response server, upon receiving a telephone call for a service provision request including designation of a service and a service provision language from a user, accesses the corresponding service provision server via a network to obtain a service name and a service. When an intermediate language is requested by designating a provided language and the intermediate language is received from the service providing server, necessary information is extracted from the intermediate language, and a dialogue sequence is created using the dialogue sequence creating means based on the information. Generate,
The speech dialogue translation service system according to any one of claims 1 to 7, wherein the speech dialogue translation service system interacts with a user according to the dialogue sequence. 9. The voice response server determines a service providing server holding an intermediate language corresponding to a service designated by the service managing means, and accesses the corresponding service providing server. The voice dialogue translation service system according to any one of claims 1 to 8. 10. The voice response server obtains the utterance content of the user when interacting with the user according to the dialogue sequence, and substitutes the information obtained by the dialogue as a value of a language-dependent part of the intermediate language. The spoken language translation service system according to any one of claims 1 to 9, wherein the intermediate language in which the value is substituted is sent to the service providing server. 11. When the service providing server receives an intermediate language request from the voice response server in the previous term, if the service providing server holds an intermediate language corresponding to the requested service name and service providing language, the service providing server uses the If the setting language of the intermediate language corresponding to the requested service name is different from the requested service providing language to the response server, the translation server requests the intermediate language corresponding to the requested service name. 11. The voice interactive translation service system according to claim 1, wherein an intermediate language corresponding to a service providing language is created and then sent to the voice response server. 12. When the service providing server creates an intermediate language corresponding to a requested service providing language from an intermediate language corresponding to a requested service name, the created intermediate language is a value in an original setting language. 12. The spoken dialogue translation service system according to claim 1, wherein both the translated value and the translated value are described separately. 13. The service providing server, when receiving the intermediate language from the voice response server in the previous term, provides the service provider with the intermediate language.
2. The speech dialogue translation service system according to any one of 2 above. 14. The service providing server, when providing information of a user to a service provider, a part depending on an intermediate language is composed of an original setting language and a value of a translated setting language. In this case, the speech dialogue translation service system according to any one of claims 1 to 13, wherein a part corresponding to the original set language is selected and provided to the service provider. 15. The service providing server, when sending an intermediate language to the voice response server, also sends a dialogue sequence generation rule corresponding to a designated language, and the dialogue sequence creating means is configured to send the dialogue sequence generating means from the service providing server. 15. The spoken dialogue translation service system according to claim 1, wherein the dialogue sequence generation rule is received, and the received dialogue sequence generation rule is used when the dialogue sequence is generated. 16. The voice response server also has a function of the service providing server.
~ The voice interactive translation service system according to any one of claims 15 to 16.