JPH0333796A - Interactive system - Google Patents

Abstract

PURPOSE:To give a feeling as if an interaction with a main is executed as possible by generating a random number according to the number of an interaction number part to indicate the number of interaction groups composed of two abnormal interactive sentences having the same meaning and different expression to a random selecting part, selecting the interactive sentence according to the random number from the interaction groups and outputting it. CONSTITUTION:By the confirmation result of a voice as the instructions of the equipment confirmation, control, etc., of a user, the interaction of the next inquiry is decided. To the interaction of the inquiry, the random number is generated by a random selecting part 13 according to the number of an interaction kind number part 12 to indicate the number of interaction groups 11 to collect interactive sentences or phrases having the same meaning and different expression, the interactive sentence or phrase according to the random number is selected from the interaction groups 11, and the voice file is synthesized as a response voice. By changing the response from a machine at random in such a way, a natural feeling as if the interaction with the man is executed as possible though an opponent is the machine from the position of the user can be given.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention makes a machine ask a question by voice or text display that requests a response from a human, or responds to a request from a human by a voice or text display from the machine. It is related to a dialogue system used for.

2. Description of the Related Art In recent years, the use of computer machines has led to a variety of systems in which machines and humans interact by voice or text, such as bank cash dispensers, automatic transfer machines, and airline automatic seat reservation systems by telephone. It has been put into practical use.

Conventionally, in this type of dialogue system, the dialogue text of a question that requests a response from Iruma, such as the OR of a bank's cash dispenser, the message on the T-display, or the dialogue voice of an automatic seat reservation system, is used to respond to the same request. It is always the same dialogue sentence or phrase.

Problems to be Solved by the Invention Even if only one dialogue sentence or phrase is prepared from the machine side, as in the conventional machine-human dialogue system described above, if the dialogue system is infrequently used,
It's not that difficult to use. However, if the dialogue system is frequently used and is used repeatedly, the dialogue sentences or phrases may become boring, and for first-time users, a friendly machine response may be all that is needed. However, for those who are accustomed to using the machine, the repeated responses of the machine using the same dialogue sentences or phrases can be troublesome.
It feels unnecessary. In order to eliminate this bothersome response, it is possible to store information such as the number of past uses and the period until the last use in a database in order to provide different responses depending on the frequency of use, but this is not practical. do not have.

The present invention solves the above-mentioned conventional problems, and by randomly changing the responses from the machine, from the user's perspective, it is possible to interact with a human being as much as possible, even though the other party is a machine. To provide a dialogue system that can give a natural feeling, and also to add an adverb or an adverbial phrase when a machine repeatedly issues a request to give an even more natural feeling to the dialogue. The purpose is to provide a dialogue system that allows users to

Means for Solving the Problems The technical solution of the present invention for achieving the above object is as follows:
A dialogue group consisting of two or more dialogue sentences or phrases with the same meaning but different phrasing as a question from a machine that requests a human response or a machine response to a human request, and a dialogue group consisting of two or more dialogue sentences or phrases that have the same meaning but are worded differently. A random number part that indicates the number of dialogue sentences or words and a random number corresponding to the number of dialogue parts is generated, and a dialogue sentence or phrase from the dialogue group is selected and output according to the random number. It is equipped with a selection section.

In addition, a re-request adverb group consisting of two or more adverbs or adverbial phrases indicating re-request, an adverb number section indicating the number of adverbs or adverb phrases in this re-request adverb group, and the number of this adverb number section. a random selection unit that generates a random number according to the response, selects an adverb or an adverbial phrase from the above-mentioned re-requested adverb group according to the random number, and outputs it by adding it to the dialogue sentence or phrase selected as above. It is prepared.

action The present invention has the following effects due to the above configuration.

As a question from a machine that requests a human response, or a machine response to a human request, two or more dialogue sentences with the same meaning but different phrasing, or two or more dialogues consisting of words and phrases with different phrasing. Generate a random number in a random selection section according to the number of dialogues indicating the number of dialogue groups consisting of sentences or phrases, select a dialogue sentence or phrase according to the random number from the dialogue group, and output it. can do.

In addition, when the same request as the previous one is repeatedly omitted from the machine, a random number is randomly selected according to the number of numerical parts indicating the number of re-request adverb groups consisting of two or more adverbs or adverbial phrases indicating re-request. The adverb or adverb phrase corresponding to the random number can be selected from the group of re-required adverbs, added to the dialogue sentence or phrase selected as described above, and output.

In this way, the questions or responses from the machine can be changed randomly.

Examples Examples of the present invention will be described below with reference to the drawings.

First, an audio telecontrol system will be described as an example of implementing the dialogue system of the present invention.

FIG. 3 is an explanatory diagram of the audio telecontrol system. The voice telecontrol system is a system that uses a telephone line to control home appliances (air conditioners, lights, electric locks, etc.) from outside the home, etc., and uses a synthesized voice 8 from the dialog section 7 to request Iruma's response. On the other hand, the user 1 generates a command voice 2 to the machine from the receiver. That audio 2
The machine recognizes 3 and passes the result 4 to the control 6 side. Based on the results, the controller 6 side controls the equipment. Then, the next audio file 6 that requires a human response is selected.

The dialogue section 7 receives the audio file 6, synthesizes and outputs the dialogue sound of the question. Then, the user 1 responds to the synthesized voice 8. Repeat the above operations to check the device status and control the device.

The dialog system of the present invention includes the control 60 in FIG.
It corresponds to the part.

FIG. 1 shows a dialogue system in a first embodiment of the present invention, and is a block diagram for selecting dialogue sentences.

In FIG. 1, reference numeral 11 is a dialogue group, in which two or more dialogues that have the same meaning but are worded differently (in the illustrated example An audio file containing 6) dialogue sentences or phrases is prepared. Reference numeral 12 denotes a dialogue number section indicating the number of dialogue sentences or phrases in the dialogue group 11, and 5 is shown in the illustrated example. Reference numeral 13 denotes a random selection section that generates a random number according to the number of dialogues number section 12, and randomly selects and outputs a dialogue sentence or phrase in the dialogue group 11 according to the random number.

As explained in FIG. 3, the next question dialogue is determined by the recognition result of the user 1's voice as a command for checking/controlling equipment, etc. For the dialog of the question, a random number is generated in the random selection section 13 according to the number in the dialog type number section 12 that indicates the number of dialog groups 110 that have the same meaning but different phrasing or phrases. , a dialogue sentence or phrase corresponding to the random number is selected from the dialogue group 11, and the audio file is synthesized as a response audio.

In a voice telecontrol system, examples of dialogue sentences that request the name of the device you want to check in order to check the current status of the device are ``Do you want to check something?'' and ``
Audio files such as "Please tell me the name of the device you want to confirm" and "Please tell me the name of the device you want to confirm" are prepared, and these audio files are randomly selected and synthesized as described above.

In addition, even when selecting dialogue sentences or phrases at random, by always using the same expression for the first time the system is used, people who are not accustomed to using the system can easily understand that the expression changes each time they use the system. Since there is no such thing, there is an advantage that it does not give a feeling of anxiety.

Furthermore, if there are many devices to be checked, the process of deciding what to check will be repeated several times. When we think about when a person repeatedly asks a question, we need to not only randomly synthesize two or more types of question expressions, but also use adverbs that indicate a repeat request such as "next" or "else". Tell a story by adding adverb phrases. Therefore, in the second embodiment of the present invention, adverbs and adverbial phrases indicating several types of re-requests that match the expression are randomly selected and added to the beginning of the selected interrogative dialogue sentence or word phrase. The response voice is synthesized by voice.

FIG. 2 shows a dialog system in a second embodiment of the present invention, and is a block diagram for selecting re-required adverbs. In FIG. 2, 21 is a re-request adverb group, and an audio file is prepared in which two or more (three in the illustrated example) adverbs or adverb phrases indicating re-request are collected. Reference numeral 22 denotes an adverb count section indicating the number of adverbs or adverbial phrases in the re-request adverb group, which is 3 in the illustrated example. 23 is adverb number part 2
2, and randomly selects an adverb or an adverb phrase from the re-request adverb group 21 according to the random number.

Therefore, the random selection section 23 generates a random number according to the number in the adverb number section 22, selects the adverb or adverbial phrase in the adverb group 21 corresponding to that number, and selects the adverb or adverbial phrase from the adverb group 21, adds an adverbial phrase to the beginning of the dialog sentence selected in FIG. 1, and synthesizes the audio file as dialog speech.

In an audio telecontrol system, when an adverb or an adverbial phrase such as "next" or "in addition" is added to the dialogue audio that confirms the current status of the three types of equipment mentioned above, and when it is not added. The following is an example of a change in a dialogue voice randomly selected.

→ “Do you want to check something else?” → “What other device names do you want to check?” → “Do you want to check something next?” → “Please tell us the name of the device you want to check.” → “What other device names do you want to check?” ->"Please tell us the name of the device you want to check next."-> In the above example, we explained the case where the machine responds by voice synthesis, but the machine's response is not limited to voice synthesis. It can also be applied to menusage on a display, such as the seven menusage for a meaningless command input in an adventure game.

Effects of the Invention As described above, according to the present invention, two or more dialogues that have the same meaning but different wording can be used as a question from the machine that requests Iruma's response, or as a response from the machine to Iruma's request. Generate a random number in a random selection section according to the number of dialogues indicating the number of dialogue groups consisting of sentences or phrases, select a dialogue sentence or phrase according to the random number from the dialogue group, and output it. This makes it possible to give the feeling of interacting with a human being as much as possible.

In addition, when the same request as the previous one is repeatedly issued from the machine, the random selection unit randomly selects a random number according to the number of numerical parts indicating the number of re-request adverb groups consisting of two or more adverbs or adverbial phrases indicating re-request. , select an adverb, an adverbial phrase corresponding to the random number from the above-mentioned re-required adverb group, and output it by adding it to the dialogue sentence or word phrase selected as above. I can do it.

In this way, since the questions or responses from the machine can be changed randomly, it is possible to provide an even more natural feeling of dialogue when the machine repeatedly issues requests.

[Brief explanation of drawings]

FIG. 1 shows a dialogue system according to a first embodiment of the present invention, and is a block diagram for selecting dialogue sentences. FIG. FIG. 3 is an explanatory diagram of an audio telecontrol system implementing the present invention. 11... Dialogue group, 12... Dialogue number part, 13...
Random selection part, 21... Re-request adverb group, 22...
Adverb number part, 23...Random selection part.

Claims (2)

[Claims] (1) A dialogue group consisting of two or more dialogue sentences or phrases that have the same meaning but are worded differently, as a question from a machine that requests a human response, or a machine response to a human request. Dialogue sentences in dialogue groups,
Alternatively, a random selection section that generates a dialogue number part indicating the number of words and a random number according to the number of this dialogue number part, and selects and outputs a dialogue sentence or a phrase from the dialogue group according to the random number. A dialogue system characterized by: (2) A re-request adverb group consisting of two or more adverbs or adverbial phrases indicating re-request, an adverb number section indicating the number of adverbs or adverb phrases in this re-request adverb group, and the number of this adverb number section A random selection unit is provided, which generates a random number according to the random number, selects an adverb or an adverbial phrase from the above-mentioned re-requested adverb group according to the random number, and outputs the selected adverb or phrase added to the selected dialogue sentence or word phrase. The dialogue system according to claim 1, characterized in that: