JP2004355226A - Inter-different language dialog equipment and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To attain inter-different language intention understanding in a practical level by using a technology such as voice recognition or machine translation in a current technology level. <P>SOLUTION: This inter-different language dialog equipment is provided with a dialog processing means to execute any processing of first processing to decide response contents to a first interlocutor by referring to preliminarily stored dialog scenario, and to transmit the response contents through a first terminal part in first language by using a stylized dialog sentence designated for every response contents, second processing to transmit a sentence where a conversion result obtained by converting an utterance understanding result being the result of utterance understanding into second language different from the first language is combined with the stylized dialog sentence through a second terminal part different from the first terminal part to a second interlocutor different from the first interlocutor and third processing to end the dialog. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a cross-language dialogue apparatus and a cross-language cross-talking method that enables communication between interlocutors using different languages in each of a plurality of terminal units.
[0002]
[Prior art]
The conventional interlingual dialogue apparatus, after performing speech recognition of all the content uttered by one speaker in a predetermined language, converts all of the speech recognition results into another language by machine translation, This is a device that converts the speech into a speech by a speech synthesis device or the like and transmits it (for example, see Non-Patent Document 1)
Non-Patent Document 1 describes in detail the technology of an automatic translation telephone which is a cross-language dialogue system.
[0003]
[Non-patent document 1]
"International Telecommunications Research Institute", "Automatic Translation Telephone" Ohmsha, ATR Advanced Technology Series, January 1994 [0004]
[Problems to be solved by the invention]
However, in the conventional interlingual dialogue apparatus, since the performance of speech recognition is not sufficient, a speech recognition result different from the input utterance content is often obtained. In addition, since the performance of machine translation is not sufficient, the result often translates into another language different from the utterance content.
[0005]
Under such circumstances, it is often not possible to accurately convert the content uttered by the speaker into another language, and the conventional technology allows communication in different languages and conversion of utterances into different languages. Is difficult to achieve on a practical level.
[0006]
The present invention provides a cross-language dialogue apparatus and a cross-language dialogue method that can communicate at a practical level when communicating in different languages using the current technology-level technologies such as speech recognition and machine translation. It is intended to provide.
[0007]
In addition, the present invention provides a machine translation apparatus and a machine translation method that can be used at a practical level when machine translation into a different language is performed by using current technology-level techniques such as speech recognition and machine translation. The purpose is.
[0008]
[Means for Solving the Problems]
The present invention determines a response content to a first interlocutor by referring to a dialog scenario stored in advance, uses a fixed dialogue sentence specified for each response content, and uses a first terminal unit. A first process of transmitting the response content in a first language via a first language, and a conversion result obtained by converting an utterance understanding result, which is a result of the utterance understanding, into a second language different from the first language. And a second process of transmitting a sentence combining the fixed dialogue sentence to a second interlocutor different from the first interlocutor via a second terminal different from the first terminal. , A cross-language dialogue device having a dialogue processing means for executing any of the third processes for ending the dialogue.
[0009]
Embodiments and Examples of the Invention
FIG. 1 is a block diagram showing a cross-language dialogue apparatus 100 according to a first embodiment of the present invention.
[0010]
The interlingual dialogue device 100 includes a terminal unit 10, a terminal unit 20, a dialogue control unit 30, a dialogue scenario storage unit 40, a fixed form dialogue data storage unit 50, and a syntax pattern storage unit 60. I have.
[0011]
The terminal unit 10 is connected to a voice recognition unit 11, and the terminal unit 20 is connected to a voice recognition unit 21.
[0012]
The dialog control unit 30 can refer to the dialog scenario storage unit 40, the fixed-form dialogue data storage unit 50, and the syntax pattern storage unit 60, and is connected to the machine translation unit 31. Further, the dialogue control unit 30 refers to the dialogue scenario storage unit 40, requests the utterance of the interlocutor A associated with the terminal unit 10 via the dialogue terminal unit 10, or obtains the utterance from the terminal unit 20 via the terminal unit 20. Ask the interlocutor B related to 20 to speak.
[0013]
The interlocutor whose utterance is requested inputs the utterance via the terminal unit 10 and obtains a speech recognition result for the utterance using the speech recognition unit 21 coupled to the terminal unit 10.
[0014]
The dialogue control unit 30 understands the utterance by referring to the syntax pattern storage unit 60 based on the result of the speech recognition. Then, by referring to the dialogue scenario storage unit 40, the “next processing” corresponding to the utterance understanding result is determined and executed.
[0015]
The option of the “next processing” in the dialog control unit 30 is any one of the first processing, the second processing, and the third processing.
[0016]
In the “first processing”, voice synthesis is performed in the language handled by the terminal unit 10 of the interlocutor A via the terminal unit 10 of the interlocutor A, and the pre-stored dialog scenario is referred to. Determines the contents of a response to the first interlocutor A, and uses a fixed dialogue sentence (described in the fixed-form dialogue data storage unit 50) specified for each response content, and uses the first terminal unit. This is a process of transmitting the response content in a first language through the user 10 and asking the interlocutor A for the next utterance.
[0017]
The “second process” is to convert a sentence obtained by combining a conversion result obtained by converting an utterance understanding result, which is a result of the utterance understanding, into a second language different from the first language, and a fixed dialogue sentence. This is a process of transmitting to a second interlocutor B different from the first interlocutor A via a second terminal unit 20 different from the first terminal unit.
[0018]
The “third process” is a process for ending the conversation.
[0019]
Until the dialog control unit 30 determines that the next process is the end of the dialog, the dialogue person A who has been asked to speak inputs the speech through the terminal unit 10 of the dialogue person A side, and Continue.
[0020]
In the utterance input at the terminal unit 10, not only voice recognition, but also a method in which an interlocutor inputs characters using a character recognition device and the system recognizes characters may be used together. The method in which the unit 10 accepts the message as it is may be used together.
[0021]
In the transmission of the utterance content using the fixed dialogue sentence in the terminal unit 10, in addition to the method of outputting the voice by the voice synthesizer, the method of screen display or the voice output and the screen display by the voice synthesizer are used in combination. A method may be adopted.
[0022]
The syntax pattern storage unit 60 is basically the same as the syntax pattern storage unit in the conventional example, but differs from the conventional syntax pattern storage unit in that a “question word” is optionally specified.
[0023]
FIG. 2 is a diagram illustrating an example of the syntax pattern stored in the syntax pattern storage unit 60 in the above embodiment.
[0024]
The syntax pattern storage unit 60 is configured in a table format having key phrase and essential element items. When an input utterance includes a key phrase, a required element is described as a required word. Is shown in
[0025]
The “key phrase” is a phrase corresponding to a keyword, and is composed of a noun or a noun phrase. The “essential element” is an element such as a subject and an object that is necessary for the predetermined sentence to constitute a complete sentence in the predetermined sentence.
[0026]
For example, in the example of the syntax pattern shown in FIG. 2, when a key phrase (phrase) of "I want you to connect" is included in the utterance, "he" or "to" is added after the key phrase. This indicates that the following phrase and the phrase following "" are required as essential elements.
[0027]
Further, in the example shown in FIG. 2, the question words corresponding to the essential elements are described in parentheses in the column of the essential elements. For example, the question word corresponding to “to N1” is “Which”. Is shown. Note that “N” is a noun or a noun phrase.
[0028]
The syntax pattern storage unit 60 stores the words and phrases of the phrase from an external dictionary or the like by defining the part of speech of an essential element such as “noun + to” or “N1” having the meaning of “person”. By defining the meaning, conditions for defining other essential elements may be described.
[0029]
The question words corresponding to the essential elements may be described in advance as in the present embodiment, or may be automatically selected based on the semantic restrictions of the essential elements.
[0030]
Further, the syntax pattern storage unit 60 may provide a syntax pattern by preparing in advance according to the content of the dialogue to be assumed, and may use a dictionary or a rule used in the machine translation system. The syntactic pattern may be provided by itself or by using a modification thereof.
[0031]
FIG. 3 is a diagram illustrating an example of the dialog scenario stored in the dialog scenario storage unit 40 in the above embodiment.
[0032]
The dialogue scenario storage unit 40 stores a dialogue scenario, and the “dialogistic scenario” is composed of a tabular scenario having items of a dialogue state, an utterance understanding result, a response content, a response destination, and a next dialogue state. . In other words, the dialogue scenario starts from the dialogue state STATE0 shown in FIG. 3 and progresses to the dialogue state STATE9 until the dialogue state continues. This is a scenario in which the terminal unit specifies which interlocutor is the response destination, responds with what kind of response content, and as a result, transitions to the next conversation state.
[0033]
FIG. 4 is a diagram illustrating an example of the fixed form dialogue data stored in the fixed form dialogue data storage unit 50 in the above-described embodiment. FIG. 7 is a diagram illustrating an example of fixed form dialogue data in which a fixed form dialogue sentence is prepared.
[0034]
When a dialogue sentence such as Japanese or English is specifically generated based on the response content, a fixed dialogue corresponding to the response content specified by the dialogue scenario shown in FIG. It obtains from the data and generates a dialogue sentence using the found fixed form dialogue.
[0035]
In the “Japanese fixed dialogue sentence” in the example of the fixed dialogue data shown in FIG. 4, in the <utterance understanding content>, the utterance understanding content obtained up to the processing is embedded, and the <missing essential elements are questioned. In the utterance comprehension contents replaced with the lyric>, a sentence that combines the utterance comprehension contents obtained so far and an expression that asks for the essential element missing from the key phrase is embedded, so that the fixed dialogue sentence is Generate.
[0036]
Here, instead of referring to the standard dialogue sentence data, a conventional technique of natural language sentence generation may be used to generate a dialogue sentence such as Japanese or English based on the response content.
[0037]
In the following, in the interlingual dialogue apparatus 100, a case will be described in which the interlocutor A interacts in Japanese through the terminal unit 10 and the interlocutor B receives the dialogue in English through the terminal unit 20. I do.
[0038]
Next, in the above-described embodiment, in accordance with the example of the syntax pattern shown in FIG. 2, the example of the dialogue scenario shown in FIG. 3, and the example of the typical dialogue sentence data shown in FIG. The following describes the operation of a dialogue between different languages, in which a request is made in English and the interlocutor B receives a request in English via the terminal unit 20.
[0039]
FIG. 5 is a diagram showing an example of the syntax pattern shown in FIG. 2, the example of the dialogue scenario shown in FIG. 3, and the example of the typical dialogue data shown in FIG. FIG. 10 is a diagram illustrating an example of a dialogue between different languages, in which a request is made in Japanese via a terminal unit 20 and a request is made in English via a terminal unit 20.
[0040]
FIG. 6 is a flowchart showing the overall processing in this embodiment.
[0041]
The dialogue is started from a dialogue state STATE0 (see FIG. 3), which is an initial state, and the dialogue control unit 30 sends a response content shown in FIG. To output a speech request. From the response content of the utterance request, referring to the Japanese standard dialogue sentence data shown in FIG. 4, respond to the terminal unit 10 with a Japanese sentence "Please tell us what you want to tell the other party," When the "dialogue state" is designated, the state transits to the dialogue state STATE1 (S1).
[0042]
In the example of the dialog shown in FIG. 5, in response to the response "Please tell us what you want to tell the other party," Interactor A speaks, "I want you to connect the phone a little." Is also obtained with the same sentence (S3).
[0043]
For this utterance, a key phrase is searched from the example of the syntax pattern shown in FIG. In this utterance, “connect”, which is underlined in FIG. 5, is searched for as a key phrase, and the essential elements “to N1 (?)” And “to N2” are obtained.
[0044]
Further, in the sentence of the speech recognition result, there is “telephone” (double line in FIG. 5) corresponding to “N2”, and there is no phrase corresponding to the essential element “to N1”. It is not specified whether or not to make a call). As a result, referring to FIG. 3, the utterance understanding result is "there is a key phrase, but essential elements for the key phrase are insufficient" (S4).
[0045]
In the example shown in FIG. 5, for example, each of the words constituting "I want to connect a phone call" spoken by the interlocutor A is compared with the words registered in a predetermined key phrase syntax pattern. If there is a match, the matched word is recognized as a key phrase. By doing so, "I want you to connect" is recognized as a key phrase.
[0046]
On the other hand, the system generates a response sentence of the system by referring to the example of the fixed-form dialogue data shown in FIG. . The standard Japanese dialogue sentence of "Requirement for missing essential elements" is "Is the utterance understanding content in which missing essential elements are replaced with question words?", And <Insufficient essential elements are replaced with question words. The sentence generated from the utterance understanding content up to that point is embedded in the “utterance understanding content>” portion.
[0047]
Here, the above-mentioned “embedded sentence” is obtained from the fixed-form dialogue sentence data shown in FIG. 4 by using the question word “to which” for the missing essential element “to N1 / to”, the remaining element “telephone”, and the key phrase “ Please connect them ".
[0048]
In other words, the response sentence here is "Which phone do you want to connect to?" Further, according to FIG. 3, the state transits from the conversation state STATE1 to the next conversation state STATE2 (S1).
[0049]
Next, as shown in FIG. 5, it is assumed that the interlocutor A replies "Please go to Room 204" and the speech recognition result incorrectly recognizes "Room 205" (S3). The speech recognition result includes “to room 205” (the double line in FIG. 5) corresponding to the essential element “to N1 / to” that was missing in the previous state. Thus, from FIG. 3, the utterance understanding result becomes “there is an essential element of the previous utterance” (S4).
[0050]
In response to this, the Japanese standard response sentence of the “confirmation of utterance content” is referred to as an example of the typical dialogue sentence data shown in FIG. ?? "is generated from the system. The sentence generated based on the utterance understanding content up to that time is embedded in the <utterance understanding content> part.
[0051]
Here, the sentence to be embedded is a sentence "I want you to connect the telephone to Room 205" in which the essential elements "To room 205", "Call me" and the key phrase "I want you to connect" are connected.
[0052]
Therefore, the response sentence here is "Do you want to tell me that you want to connect to room 205?", And responds to this message to the interlocutor A through the terminal unit 10. Further, according to FIG. 3, the state transits from the conversation state STATE2 to the next conversation state STATE3 (S1).
[0053]
Next, as shown in FIG. 5, the interlocutor A answers "I'm going to room 204" and recognizes that the speech recognition result was obtained without error this time (S3). This speech recognition result includes “to room 204” (double line in FIG. 5) corresponding to the already required essential element “to N1”.
[0054]
As a result, based on the example of the dialogue scenario shown in FIG. 3, the utterance understanding result is determined to be “correction of the understanding contents” (S4), and for the essential element “to N1”, to “room 204”. "Is updated. In response to this, the Japanese standard response sentence of the “confirmation of utterance content” is referred to as an example of the typical dialogue sentence data shown in FIG. ?? "is generated from the system.
[0055]
The sentence generated based on the utterance understanding contents up to that time is embedded in the <utterance understanding contents> part. Here, the sentence to be embedded is a sentence "I want you to connect the telephone to Room 204" in which the essential elements "To room 204", "Call me" and the key phrase "I want you to connect" are connected. In other words, the response sentence here is "Do you want to inform that you want to connect to room 204?", And this response is sent to the interlocutor A via the terminal unit 10. Further, according to FIG. 3, the self-transition is made from the conversation state STATE3 to the next conversation state STATE3 (correction) (S1).
[0056]
The dialogue control unit 30 understands that the next utterance of the interlocutor A is "yes" (S3) as "yes" (S4), and in accordance with the response contents and the designation of the response destination in FIG. "Transmit the utterance contents" to the terminal unit 20 of the interlocutor B, and further, respond to the "utterance request" to the terminal unit 10 of the interlocutor A (S1).
[0057]
The language on the terminal unit 20 side is English, and the standard English response sentence of the response content “transmit the utterance content” is “He / Say sails,“ <utterance understanding content> ”from FIG. In the <speech comprehension content> portion of the response sentence, the above-mentioned converted to English of "I want you to connect a telephone to room 204" is embedded. This embedded sentence is obtained as a result of converting from Japanese to English by the machine translation unit 31 and is obtained as “I world like you to put me through to Room 204”, and the response sentence “He / Saysay,“ I world ” It transmits like like to put me through to Room 204 ″ ”to the terminal unit 20 of the interlocutor B. A conventional machine translation technique may be used as the machine translation used in the machine translation unit 31.
[0058]
In addition, a response is made to the terminal unit 10 of the interlocutor A with a Japanese standard dialogue sentence "Please tell us what you want to tell the other party" corresponding to the response content "request for utterance" to prompt the next utterance. After that, the dialogue state transitions to STATE1 according to the designation of the next dialogue state in FIG. 3 (S1).
[0059]
As shown in FIG. 5, from the speech recognition result of the next utterance of the interlocutor A, "OK, please end" (S3), the system understands the utterance as "End" from FIG. 3 (S4), and the terminal unit The template dialogue sentence "skilled" for "accept" is transmitted to the interlocutor A via 10 (S1). Here, since the next dialogue state is STATE 9, that is, the end state (S2), the dialogue ends.
[0060]
By the method described above, speech recognition errors are corrected interactively, and words required as input sentences for machine translation are obtained from the interlocutor, so that the utterance content processed in the dialogue system becomes more accurate, and as a result, Thus, smoother communication between different languages can be realized.
[0061]
Note that the examples shown in FIGS. 5 and 6 can be applied to the case of a telephone company in a general company and the like other than the present embodiment.
[0062]
The second embodiment of the present invention is an embodiment in which the interactive system of the above embodiment is used in combination in two directions.
[0063]
That is, the dialogue system according to the above-described embodiment uses the terminal unit 20 to request the utterance of the user B via the terminal unit 20 and reply to the dialogue in English. Applicable when receiving in English. As described above, by using the inter-language dialogue device 100 in two directions simultaneously, two-way communication between two different languages is possible.
[0064]
FIG. 7 is a block diagram showing a machine translation apparatus 200 according to a third embodiment of the present invention.
[0065]
The machine translation device 200 is obtained by removing the terminal unit 20 and the speech recognition unit 21 from the interlingual interaction device 100.
[0066]
That is, the machine translation device 200 may be configured so that the dialogue control unit 30 requests the utterance of the interlocutor A via the terminal unit 10 and receives the dialogue content returned in the first language in English via the terminal unit 10. It is. The machine translation device 200 need only use the same “first terminal unit 10” for all the response destinations of the dialog scenario shown in FIG. 3 in the interlingual dialog device 100.
[0067]
That is, the machine translation device 200 is a machine translation device that converts the utterance content from the first language to the second language, and converts the sentence spoken by the first interlocutor through the terminal unit in the first language. Referencing speech understanding means for understanding the contents of the utterance sentence by recognizing the speech and checking the content of the utterance sentence by comparing the pattern of the utterance sentence with a pre-stored syntax pattern. The first responder determines a response content to the first interlocutor, uses a fixed dialogue sentence specified for each response content, and transmits the response content in a first language via the terminal unit. The terminal unit by combining the conversion result obtained by converting the utterance understanding result into the second language different from the first language with the conversion result obtained by machine translation and the fixed dialogue sentence Through it, is an example of a machine translation apparatus and a dialogue processing means for executing said first second processing for transmitting the interlocutor, any of the third process of ending the conversation.
[0068]
The above embodiment is an example in which only one key phrase is included in one utterance. In the case of an example in which a plurality of key phrases are included, a request processing of an essential element is performed for each key phrase. The dialog scenario in the dialog scenario storage section 40 may be described so that is repeated.
[0069]
Furthermore, the above-described embodiment is not limited to the above-described embodiment, and may change and apply a language type, a character code, a dialog scenario, a syntax pattern, a fixed dialogue sentence, and the like.
[0070]
【The invention's effect】
According to the first and second aspects of the present invention, when a computer assists a dialogue between different languages, it is possible to interactively correct a speech recognition error with respect to an utterance of an interlocutor, The necessary words can be obtained interactively as input sentences, and by controlling the dialogue while checking the utterance comprehension results, even when using the current technology-level speech recognition and machine translation, The effect is that communication can be realized.
[0071]
According to the third and fourth aspects of the invention, when machine translation into a different language is performed using a technology such as speech recognition or machine translation at the current technical level, an effect can be obtained at a practical level.
[Brief description of the drawings]
FIG. 1 is a diagram showing a cross-language interactive device 100 according to a first embodiment of the present invention.
FIG. 2 is a diagram illustrating an example of a syntax pattern stored in a syntax pattern storage unit 60 in the embodiment.
FIG. 3 is a diagram showing an example of a dialog scenario stored in a dialog scenario storage unit 40 in the embodiment.
FIG. 4 is a diagram showing an example of fixed form dialogue data stored in a fixed form dialogue data storage unit 50 in the embodiment, and corresponds to a response content in the fixed form dialogue data for each language. FIG. 7 is a diagram illustrating an example of fixed form dialogue data in which a fixed form dialogue sentence is prepared.
5 is a diagram showing an example of a syntax pattern shown in FIG. 2, an example of a dialog scenario shown in FIG. 3, and an example of fixed form dialogue data shown in FIG. FIG. 13 is a diagram showing an example of a dialogue between different languages, in which a request is made in English and a request is made in English by the interlocutor B via the terminal unit 20 via the terminal unit 20.
FIG. 6 is a flowchart illustrating overall processing in the present embodiment.
FIG. 7 is a block diagram illustrating a machine translation apparatus 200 according to a third embodiment of the present invention.
[Explanation of symbols]
100: Interlanguage dialogue device,
10, 20 ... terminal unit,
11, 21 ... voice recognition unit,
30: Dialogue control unit,
40: Dialogue scenario storage unit
50: fixed form dialogue data storage unit
60 ... Syntax pattern storage unit
200: Machine translation device.

Claims (4)

In a cross-language dialogue device that interacts through a plurality of terminals and uses a different language in each terminal,
A first interlocutor voice-recognizes a sentence uttered in a first language via a first terminal unit, and collates with a syntax pattern in which patterns of components of the uttered sentence are described and stored in advance. Utterance understanding means for understanding the contents of the utterance sentence;
The contents of a response to the first interlocutor are determined by referring to a dialog scenario stored in advance, and a fixed dialogue sentence specified for each response is used to determine the response through the first terminal unit. A first process of transmitting the response content in a first language, a conversion result obtained by converting an utterance understanding result that is a result of the utterance understanding into a second language different from the first language, A second process of transmitting a sentence combining the dialogue sentence to a second interlocutor different from the first interlocutor via a second terminal different from the first terminal; Interactive processing means for executing any of the third processing for ending the processing;
A cross-language dialogue device, comprising: In a cross-language dialogue method of interacting through a plurality of terminals and using a different language at each terminal,
A first interlocutor voice-recognizes a sentence uttered in a first language via a first terminal unit, and collates with a syntax pattern in which patterns of components of the uttered sentence are described and stored in advance. An utterance understanding step for understanding the contents of the utterance sentence;
The contents of a response to the first interlocutor are determined by referring to a dialog scenario stored in advance, and a fixed dialogue sentence specified for each response is used to determine the response through the first terminal unit. A first process of transmitting the response content in a first language, a conversion result obtained by converting an utterance understanding result that is a result of the utterance understanding into a second language different from the first language, A second process of transmitting a sentence combining the dialogue sentence to a second interlocutor different from the first interlocutor via a second terminal different from the first terminal; Performing a process of any of a third process of ending the process;
A cross-language dialogue method comprising: In a machine translator for converting utterance content from a first language to a second language,
By speech-recognizing a sentence spoken by the first interlocutor via the terminal unit in the first language, and by collating with a pre-stored syntax pattern in which a pattern of a component of the spoken sentence is described. Utterance understanding means for understanding the content of the utterance sentence;
The content of the response to the first interlocutor is determined by referring to the dialog scenario stored in advance, and the content of the response is determined via the terminal unit using a fixed dialogue sentence specified for each response content. A first processing transmitted in a first language, combining a conversion result obtained by converting the utterance understanding result into the second language different from the first language by a machine translation and a fixed dialogue sentence And a dialogue processing means for executing one of a second process for transmitting to the first interlocutor via the terminal unit and a third process for ending the dialogue;
A machine translation device comprising: In a machine translation method for converting speech content from a first language to a second language,
By speech-recognizing a sentence spoken by the first interlocutor via the terminal unit in the first language, and by collating with a pre-stored syntax pattern in which the patterns of the components of the spoken sentence are described. An utterance understanding step of understanding the contents of the utterance sentence;
The content of the response to the first interlocutor is determined by referring to the dialog scenario stored in advance, and the content of the response is determined via the terminal unit using a fixed dialogue sentence specified for each response content. A first processing transmitted in a first language, combining a conversion result obtained by converting the utterance understanding result into the second language different from the first language by a machine translation and a fixed dialogue sentence A second processing for transmitting to the first interlocutor via the terminal unit, and a third processing for ending the dialogue;
A machine translation method, comprising:

Images