JP7347990B2 - Automatic update of automatic translation machine using voice recognition using artificial intelligence, update distribution and information system - Google Patents

Description

The present invention relates to an automatic update of an automatic translation machine using voice recognition using artificial intelligence, update distribution thereof, and an information system.

BACKGROUND ART Translation devices that translate an input language into another language and output the translation are currently in widespread use (for example, see Patent Document 1 below).

Japanese Patent Application Publication No. 5-61908

Conventional translation devices have language databases for multiple languages (Japanese, English, etc.) and their translation relationships, but for example, these languages only support the standard language of each country. Even within the same country, it does not handle subtly different pronunciations (for example, accents and dialects) depending on the region.

In this way, conventional translation devices are capable of translating pre-stored pronunciations, but cannot translate other pronunciations.

SUMMARY OF THE INVENTION In view of the above technical problems, an object of the present invention is to provide a translation terminal and a translation system that make it possible to translate pronunciations that are not stored in advance.

A translation terminal according to a first invention for solving the above problems,
A user interface that performs input/output to the outside,
A control unit that analyzes audio content input from the outside via the user interface, converts the analyzed audio content into another language, and transmits a signal of the converted audio content to the user interface. ,
The control unit is characterized in that, if the audio content cannot be analyzed, the control unit executes an operation for requesting another pronunciation by transmitting a signal requesting content for another pronunciation to the user interface.

A translation terminal according to a second invention for solving the above problems,
In the translation terminal according to the first invention,
The request for different pronunciation is any one of the following: requesting accurate pronunciation, requesting pronunciation in standard language, requesting pronunciation without accent, and requesting other expressions. It is characterized by being

A translation terminal according to a third invention for solving the above problems,
In the translation terminal according to the first or second invention,
Repeating the operation of requesting another pronunciation until the control unit can analyze the audio content;
The content of requesting the different pronunciation in the different pronunciation requesting operation is characterized in that it does not overlap with the content of requesting the different pronunciation in the previous different pronunciation requesting action.

A translation terminal according to a fourth invention for solving the above problems,
In the translation terminal according to the third invention,
comprising a storage unit storing a first language database regarding multiple types of languages and their translation relationships;
When the control unit analyzes the voice content after the different pronunciation request operation is performed one or more times, the control unit determines that the voice content has the same meaning as the first voice content, and stores the voice content in the first language database. is characterized in that this judgment information is reflected.

A translation terminal according to a fifth invention for solving the above problems,
In the translation terminal according to the third invention,
a storage unit storing a first language database regarding multiple types of languages and their translation relationships;
If the audio content is analyzed after the alternate pronunciation request operation is performed one or more times, it is determined that the audio content has the same meaning as the initial audio content, and the first language database is rewritten. and a first artificial intelligence that determines whether or not to
When the first artificial intelligence determines to perform rewriting, the information of the determination is reflected in the first language database.

A translation system according to a sixth invention for solving the above problems,
A translation terminal according to the fourth or fifth invention,
a server capable of communicating with the translation terminal and storing a second language database regarding the plurality of languages and their translation relationships;
The present invention is characterized in that when the server receives the judgment information from the translation terminal, the judgment information is also reflected in the second language database.

A translation system according to a seventh invention for solving the above problems,
In the translation system according to the sixth invention,
The server includes a second artificial intelligence that determines whether to rewrite the second language database when receiving the judgment information from the translation terminal,
The second language database is characterized in that when the second artificial intelligence determines to perform rewriting, information on the determination is reflected in the second language database.

A translation system according to an eighth invention for solving the above problem,
In the translation system according to the fifth invention,
comprising a plurality of the translation terminals,
The server receives the judgment information from one of the translation terminals, and when reflecting the judgment information in the second language database, transmits the judgment information to the other translation terminal,
The other translation terminals each reflect the judgment information received from the server in the first language database.

According to the translation terminal and translation system according to the present invention, it is possible to translate pronunciations that are not stored in advance.

FIG. 1 is a block diagram illustrating the configuration of a translation terminal and a translation system according to an embodiment of the present invention. It is a flow chart explaining operation of a translation terminal and a translation system concerning an example of the present invention. FIG. 1 is a block diagram illustrating the configuration of a translation terminal and a translation system according to an embodiment of the present invention.

EMBODIMENT OF THE INVENTION Hereinafter, the translation terminal and translation system based on this invention are demonstrated using drawing in an Example.

First, the configuration of a translation terminal (translation terminal 1) and a translation system according to this embodiment will be explained using FIG. The translation system according to this embodiment includes translation terminals 1, 1', 1", ... and a server 3. The translation terminals 1, 1', 1", ... are connected to a network (LAN or Internet, etc.) 2. It is connected to the server 3 via the server 3 so as to be able to communicate with each other.

The translation terminal 1 mainly includes a display section 10, a voice input section 11, a voice output section 12, a communication section 13, a storage section 14, and a control section 15. Among these, the display unit 10, the audio input unit 11, and the audio output unit 12 are user interfaces that perform input and output to the outside.

The display unit 10 is a touch panel or the like that provides various displays and allows the user to make various settings including the output language.

The audio input unit 11 is a microphone or the like that converts audio input from the outside of the translation terminal 1 into an electrical signal and transmits the converted electrical signal to an input processing unit 21 to be described later.

The audio output unit 12 is a speaker or the like that outputs audio to the outside of the translation terminal 1 according to an electrical signal received from an output processing unit 22, which will be described later.

The communication unit 13 is a communication interface that can communicate with the server 3 via the network 2. Note that the server 3 stores a language database 3A as an initial setting, which is a database of multiple types of languages (standard language) (Japanese, English, etc.) and their translations. This language database 3A is rewritable.

The storage unit 14 is a storage that stores various information. A language database 14A, which is a database related to multiple types of languages (standard languages) and their translations, is also stored in this storage unit 14. Note that the storage unit 14 uses the communication unit 13 to send and receive “information that has rewritten the language database” to and from the server 3, so that the language database 14A and the language database 3A have the same contents.

The control unit 15 includes a selection request unit 20, an input processing unit 21, an output processing unit 22, a language analysis unit 23, a language conversion unit 24, a request setting unit 25, and a language connection unit 26, and mainly includes: Analyze the audio content input from the outside via the audio input unit 11 (user interface), convert the analyzed audio content into another language, and send the signal of the converted audio content to the audio output unit 12 (user interface). ).

The selection request unit 20 transmits to the display unit 10 content requesting various selections including output language settings.

The input processing unit 21 transmits the content received from the voice input unit 11 to a language analysis unit 23, which will be described later. On the other hand, the output processing section 22 transmits the contents received from the language conversion section 24 and the request setting section 25, which will be described later, to the audio output section 12.

The language analysis section 23 analyzes the audio content based on the signal received from the input processing section 21. The audio content analyzed by the language analysis section 23 is sent to the language conversion section 24. However, if the speech content cannot be analyzed by the language analysis section 23, a signal is sent to a request setting section 25, which will be described later.

The language conversion unit 24 converts (translates) the audio content analyzed by the language analysis unit 23 into an output language set by the user on the display unit 10 in advance.

When the request setting unit 25 receives a signal from the language analysis unit 23 (the language analysis unit 23 could not analyze the audio content), the request setting unit 25 sends a signal requesting a different pronunciation to the audio via the output processing unit 22. It is transmitted to the output unit 12 (hereinafter referred to as a separate pronunciation request operation).

Specifically, "content of requesting different pronunciation" includes "requesting accurate pronunciation," "requesting pronunciation in standard Japanese," "requesting pronunciation without an accent," and "Request other expressions."

The request setting section 25 repeatedly executes the operation of requesting different pronunciation until the language analysis section 23 can analyze the voice content. If the request setting section 25 repeats the above operation a predetermined number of times and the language analysis section 23 is unable to analyze it, it sends a message to the voice output section 12 that the analysis is impossible.

Further, it is assumed that the "contents for requesting a different pronunciation" (in the alternative pronunciation request operation) do not overlap with the previous "contents for requesting a different pronunciation" (in the alternative pronunciation request action). For example, if the previous request was to ``request accurate pronunciation,'' the next request would be ``request pronunciation in standard Japanese.''

When the language analysis unit 23 analyzes the voice content after the separate pronunciation request operation (and corresponding voice input from the outside) is performed one or more times, the language connection unit 26 analyzes the voice content that has been analyzed. It is determined that it has the same meaning as the first audio content (linking), and the determination information is transmitted to the storage section 14 and also to the server 3 via the communication section 13.

Further, the above judgment information is reflected (rewritten) in the language database 14A (first language database) of the storage unit 14 and the language database 3A (second language database) of the server 3.

Furthermore, when the language database 3A is rewritten in this way, the server 3 transmits the above judgment information (that is, the information of the rewritten language database 3A) to other translation terminals 1', 1'', etc. As a result, the above judgment information is also reflected (rewritten) in the language databases (first language databases) not shown in the other translation terminals 1', 1'', . . . . Therefore, each translation terminal 1, 1', 1'', . . . has its language database automatically updated, and maintains a state in which the same language database is stored.

Note that in this embodiment, the contents transmitted from the language conversion section 24 and the request setting section 25 are outputted as audio by the output processing section 22 and the audio output section 12, but the present invention is not limited to this. Instead, it may be displayed as character data on the display unit 10, for example.

Furthermore, in this embodiment, it is assumed that the language database 14A is automatically rewritten based on the judgment information of the language linking unit 26, but the present invention is not limited to such a configuration. Alternatively, the translation terminal 1 (and 1′, 1″, . . . ) may be equipped with an AI (Artificial Intelligence) 31 instead of the language linking unit 26.

In addition to performing the same processing as the language linking unit 26, the AI 31 determines whether or not to rewrite the language database 14A, and when it is determined to rewrite the language database 14A, the language database 14A is rewritten (as described above). (judgment information is reflected).

Further, in this embodiment, the language database 3A is automatically rewritten (via the network 2) based on the judgment information of the language linking unit 26, but the present invention is not limited to such a configuration. Instead, as shown in FIG. 3, the server 3 may be equipped with an AI 32.

That is, when the server 3 receives the above judgment information from the translation terminal 1 (by the language linking unit 26 or the AI 31), the AI 32 (or the AI 32 and the server administrator) determines whether or not to rewrite the language database 3A. However, when it is decided to rewrite, the language database 3A is rewritten (the above judgment information is reflected) and the above judgment information is transmitted to other translation terminals 1', 1'',... Good too.

The above is the explanation of the configuration of the translation terminal and translation system according to this embodiment. An example of the operation of the translation terminal and translation system according to the present embodiment configured as described above will be explained below using the flowchart of FIG. 2.

In step S1, the selection requesting section 20 causes the display section 10 to display a request for setting the output language, and the user sets the output language. Note that this step does not necessarily have to be performed every time. For example, it may be required only at the first operation, or a specific language (for example, English, etc.) may be set as an initial setting from the time of shipment, and the user may change it as desired.

In step S2, the user inputs voice to the voice input section 11. Step S3 is a step that branches depending on whether or not the audio content can be analyzed. If the analysis is possible, the process moves to step S5, and if it is not possible, the process moves to step S4.

In step S4, the request setting unit 25 requests another pronunciation (another pronunciation request operation). The contents of this request are outputted from the audio output section 12 or the display section 10 and transmitted to the user. After executing step S4, the process returns to step S2.

In step S5, the language analysis section 23 analyzes the audio content. Step S6 is a step that branches depending on whether or not step S4 (another pronunciation request operation) has been performed. If the alternative pronunciation request operation has been performed, the process advances to steps S7A and S7B, and if not, the process advances to step S7.

In step S7, the language conversion unit 24 converts the audio content into another language (the language set in step S1). In step S8, the language converted in step S7 is output from the audio output section 12 or the display section 10.

Steps S7A and S8A perform the same operations as steps S7 and S8, respectively, and will not be described here. Further, steps S7B, S8B, and S9 are operations performed in parallel with steps S7A and S8A after the audio content is analyzed in step S5.

In step S7B, the language linking unit 26 or the AI 31 determines (links) that the initial audio content and the audio content analyzed (by the language analysis unit 23) have the same meaning.

In step S8B, the language database 3A and the language database 14A in the other translation terminals 1', 1'', .
The above is an explanation of the operation of the translation system according to this embodiment.

In this way, the translation terminal and translation system according to this embodiment can translate pronunciations that are not stored in advance.

Further, by repeating the operation of requesting different pronunciations, translation can be performed more reliably. Furthermore, by transmitting and receiving information using a server, the language databases of each translation terminal can be made the same, and translation accuracy can be further improved.

INDUSTRIAL APPLICATION This invention is suitable for the automatic update by artificial intelligence of the automatic translation machine by voice recognition, its update distribution, and an information system.

1, 1', 1'' Translation terminal 2 Network 3 Server 3A Language database 10 Display unit 11 Voice input unit 12 Voice output unit 13 Communication unit 14 Storage unit 14A Language database 15 Control unit 20 Selection request unit 21 Input processing unit 22 Output processing Section 23 Language analysis section 24 Language conversion section 25 Request setting section 26 Language connection section 31, 32 AI (artificial intelligence)

Claims (5)

A user interface that performs input/output to the outside,
A control unit that analyzes audio content input from the outside via the user interface, converts the analyzed audio content into another language, and transmits a signal of the converted audio content to the user interface. ,
If the voice content cannot be analyzed, the control unit executes a different pronunciation request operation of transmitting a signal requesting a different pronunciation to the user interface;
Repeating the operation of requesting another pronunciation until the control unit can analyze the audio content;
The content of requesting the different pronunciation in the different pronunciation request action shall not overlap with the content of requesting the different pronunciation in the previous different pronunciation request action,
comprising a storage unit storing a first language database regarding multiple types of languages and their translation relationships;
When the control unit analyzes the voice content after the different pronunciation request operation is performed one or more times, the control unit determines that the voice content has the same meaning as the first voice content, and stores the voice content in the first language database. will reflect the information of the relevant judgment.
translation terminal and
a server capable of communicating with the translation terminal and storing a second language database regarding the plurality of languages and their translation relationships;
When the server receives the judgment information from the translation terminal, the judgment information is also reflected in the second language database.
A translation system characterized by: The contents of requesting different pronunciation on the translation terminal include requesting accurate pronunciation, requesting pronunciation in standard language, requesting pronunciation without accent, and requesting other expressions. One of these
The translation system according to claim 1. The translation terminal is
a storage unit storing a first language database regarding multiple types of languages and their translation relationships;
If the audio content is analyzed after the alternate pronunciation request operation is performed one or more times, it is determined that the audio content has the same meaning as the initial audio content, and the first language database is rewritten. and a first artificial intelligence that determines whether or not to
When the first artificial intelligence decides to perform rewriting, the information of the decision is reflected in the first language database.
The translation system according to claim 1 or 2. The server includes a second artificial intelligence that determines whether to rewrite the second language database when receiving the judgment information from the translation terminal,
2. The translation system according to claim 1, wherein when the second artificial intelligence determines to perform rewriting, information on the determination is reflected in the second language database. comprising a plurality of the translation terminals,
The server receives the judgment information from one of the translation terminals, and when reflecting the judgment information in the second language database, transmits the judgment information to the other translation terminal,
5. The translation system according to claim 1 , wherein the other translation terminals respectively reflect the judgment information received from the server in the first language database.

Images