Agent processing device, agent generation system, and program

Abstract

To provide an agent processing device, an agent generation system, and a program that enable an agent to be set more easily and in detail.SOLUTION: An agent processing server 1 includes: an element information reception unit 13 that receives element information corresponding to each of a plurality of elements constituting one agent from a setting terminal; a setting file creation unit 14 that stores a plurality of pieces of element information for one agent received by the element information reception unit 13 as one setting file in a setting file storage unit 25; and an agent generation unit 17 that generates one agent by applying the plurality of pieces of element information for one agent stored in the setting file, an AI engine, and a voice recognition engine.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to an agent processing device, an agent generation system, and a program.

Background Art

[0002] Conventionally, a mechanism has been disclosed in which a display device such as a display and a microphone are installed at a reception or the like, and a virtual character is displayed so that the character responds to a user on behalf of a person (for example, Patent Documents 1 and 2). In addition, a virtual character generation device has been disclosed that is devised to obtain the satisfaction of a customer user (for example, Patent Document 3).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0004] What is disclosed in the above documents is not capable of finely setting an agent that is a virtual character.

[0005] Therefore, an object of the present invention is to provide an agent processing device, an agent generation system, and a program that enable easier and finer setting of an agent.

Means for Solving the Problems

[0006] The present invention solves the above problems by the following means. The first invention is an agent processing apparatus comprising: element information receiving means for receiving element information corresponding to each of a plurality of elements constituting one agent from a setting terminal; setting storage means for storing the plurality of pieces of element information for the one agent received by the element information receiving means in a setting file; and agent generation means for generating the one agent to which the plurality of pieces of element information for the one agent stored in the setting file, an AI engine, and a speech recognition engine are applied. The second invention is an agent processing apparatus according to the agent processing apparatus of the first invention, further comprising: an element storage unit storing a plurality of pieces of the element information corresponding to each of the plurality of elements; and element candidate output means for outputting the plurality of pieces of element information to the setting terminal for each of the elements stored in the element storage unit, wherein the element information receiving means receives, from the setting terminal, the element information selected by a user of the setting terminal from the plurality of pieces of element information for each of the elements. The third invention is an agent processing apparatus according to the agent processing apparatus of the second invention, wherein the element information receiving means receives, from the setting terminal, the element information corresponding to the element input by the user of the setting terminal in the setting terminal. The fourth invention is an agent processing apparatus according to any one of the agent processing apparatuses of the first to third inventions, wherein the elements include at least appearance, voice color, personality, and knowledge, and the element information includes setting contents of the elements. The fifth invention is an agent processing apparatus according to any one of the agent processing apparatuses of the first to fourth inventions, further comprising: output destination receiving means for receiving, from the setting terminal, a designation of an output terminal for outputting the one agent to be generated; and agent output means for outputting the one agent generated by the agent generation means to the designated output terminal received by the output destination receiving means. The sixth invention is an agent processing apparatus according to the fifth invention, further comprising: an interactive reception unit that receives interactive input data via the output terminal; and an interactive output unit that associates the interactive output data generated by applying the interactive input data to the AI engine and the speech recognition engine with the agent output by the agent output unit and outputs the associated data to the output terminal. The seventh invention is an agent processing apparatus according to the sixth invention, wherein the AI engine executes a process of automatically generating a response text for the input text of the interactive input data using the element information for the one agent stored in the setting file. The eighth invention is an agent processing apparatus according to the seventh invention, wherein the speech recognition engine generates text from the interactive input data received by the interactive reception unit and executes a process of generating the interactive output data from the response text generated by the AI engine. The ninth invention is an agent processing apparatus according to any one of the first to eighth inventions, further comprising: a file reception unit that receives a selection of the setting file stored by the setting storage unit from the setting terminal; and an element information output unit that outputs the plurality of element information stored in the setting file corresponding to the reception by the file reception unit to the setting terminal. The element information reception unit receives the changed element information from the setting terminal, and the setting storage unit stores the changed element information received by the element information reception unit in the setting file. The tenth invention is an agent generation system comprising: an agent processing apparatus according to any one of the first to ninth inventions; and a setting terminal communicably connected to the agent processing apparatus. The eleventh invention is a program for causing a computer to function as an agent processing apparatus according to any one of the first to ninth inventions.

Advantages of the Invention

[0007] According to the present invention, it is possible to provide an agent processing apparatus, an agent generation system, and a program that enable easier and more detailed setting of an agent.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Embodiment for Carrying Out the Invention

[0009] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. Note that this is merely an example, and the technical scope of the present invention is not limited thereto. (Embodiment) <Agent Generation System 100> FIG. 1 is an overall schematic diagram of an agent generation system 100 according to the present embodiment.

[0010] The agent generation system 100 shown in FIG. 1 is a system including an agent processing server 1 (agent processing device), a setting terminal 4, and an output terminal 5. The agent generation system 100 generates an agent based on element information about a plurality of elements constituting the agent received by the agent processing server 1 via the setting terminal 4. Here, an agent refers to, for example, a digital character in the form of a human, an animal, a robot, etc., capable of communicating with an interlocutor (user) by voice or gesture. Also, the plurality of elements constituting the agent are items related to the appearance and inner aspect of the agent that characterize the agent.

[0011] Then, the agent generation system 100 outputs the generated agent to the designated output terminal 5 based on the designation of the output destination received by the agent processing server 1 via the setting terminal 4. As a result, the agent generation system 100 can easily generate an agent that finely reflects the preferences of the setter (user) operating the setting terminal 4 and output it to the output terminal 5. The agent processing server 1, the setting terminal 4, and the output terminal 5 are communicably connected via a communication network N. The communication network N is, for example, an Internet line or the like, and it does not matter whether it is wired or wireless. Also, at least a part of the communication network N may be a dedicated line.

[0012] <Agent Processing Server 1> FIG. 2 is a functional block diagram of the agent processing server 1 according to the present embodiment. The agent processing server 1 is a device that generates an agent based on element information about a plurality of elements constituting the agent received by the agent processing server 1 via the setting terminal 4. And the generated agent conducts a conversation using AI (artificial intelligence) for thinking. As a result, the agent processing server 1 enables communication between the agent output to the output terminal 5 and the interlocutor.

[0013] The agent processing server 1 is, for example, a server device. The agent processing server 1 may also be an information processing device such as a personal computer (PC). In addition, although the agent processing server 1 is described as one server in FIG. 1, it is not limited thereto. There is no limit to the number of hardware components constituting the agent processing server 1, and it may be configured as one or more as needed. Further, the hardware of the agent processing server 1 may be configured to include various servers such as a web server, a DB server, and an application server as needed, and may be configured with one server, or each may be configured with a separate server. Further, the agent processing server 1 may be, for example, at least partially in the cloud.

[0014] As shown in FIG. 2, the agent processing server 1 includes a control unit 10, a storage unit 20, and a communication interface unit 39. The control unit 10 is a central processing unit (CPU) that controls the entire agent processing server 1. The control unit 10 appropriately reads and executes the operating system (OS) and application programs stored in the storage unit 20, and cooperates with the above-described hardware to execute various functions.

[0015] Before explaining the control unit 10, the storage unit 20 will be explained. The storage unit 20 is a storage area such as a hard disk or a semiconductor memory element that stores programs, data, etc. necessary for the control unit 10 to execute various processes. The storage unit 20 includes a program storage unit 21, an element storage unit 23, a setter information storage unit 24, a setting file storage unit 25, an AI engine unit 31, and a voice recognition engine unit 32. The program storage unit 21 is a storage area that stores various programs. The program storage unit 21 stores an agent program 21a. The agent program 21a is a program for performing various functions described later that are executed by the control unit 10 of the agent processing server 1.

[0016] The element storage unit 23 is a storage area that stores element information for each element. Here, an element refers to an item representing each feature that constitutes an agent, and includes at least appearance, voice, personality, and knowledge regarding the appearance and inner surface of the agent. Also, element information is the content about an element. For example, in the case of appearance, it is parameter data representing each avatar image. Element information may further be associated with an element ID (Identification) for identifying the element information. Examples of each element stored in the element storage unit 23 and its content will be described separately later based on a screen example.

[0017] The setter information storage unit 24 is a storage area that stores information regarding a setter who sets an agent by selecting and / or inputting element information of each element of the agent. The setter information storage unit 24 stores, for example, by using a setter ID for identifying a setter as a key, in association with a file ID. The file ID is identification information for identifying a setting file that stores element information for each element set by the setter. The setting file storage unit 25 is a storage area that stores setting files. The setting file storage unit 25 stores element information, for example, by using the file ID of the setting file as a key. The file ID is identification information assigned to each agent. The element information may be parameter data corresponding to each element or may be an element ID.

[0018] The AI engine unit 31 is an area that stores an AI engine. The AI engine automatically generates a response text for an input text of a dialogue. The speech recognition engine unit 32 is an area that stores a speech recognition engine. The speech recognition engine generates an input text from dialogue input data that is speech data, and generates dialogue output data that is speech data from the response text.

[0019] Next, the control unit 10 will be described. The control unit 10 includes an agent setting unit 11 and an agent generation processing unit 15. The agent setting unit 11 is a control unit that receives element information of each element related to the configuration of the agent from the setting terminal 4 and creates a setting file used for generating the agent. The agent setting unit 11 includes an element candidate output unit 12, an element information reception unit 13, and a setting file creation unit 14.

[0020] The element candidate output unit 12 has a function as element candidate output means. The element candidate output unit 12 outputs a plurality of pieces of element information to the setting terminal 4 for each element stored in the element storage unit 23. The element information reception unit 13 has a function as element information reception means. The element information reception unit 13 receives element information corresponding to each of a plurality of elements from the setting terminal 4. The element information reception unit 13 may receive, from the setting terminal 4, the element information selected by the agent setter who operates the setting terminal 4 from among the plurality of pieces of element information for each element output by the element candidate output unit 12. Further, the element information reception unit 13 may receive, from the setting terminal 4, the element information corresponding to the element input by the agent setter at the setting terminal 4.

[0021] The setting file creation unit 14 has a function as setting storage means. The setting file creation unit 14 creates a setting file for the agent in the setting file storage unit 25 and stores the plurality of pieces of element information received by the element information reception unit 13 in the created setting file. By the processing of the setting file creation unit 14, the element information of each element necessary for generating one agent is stored in one setting file.

[0022] The agent generation processing unit 15 is a control unit that generates an agent using the setting file, outputs it to the output terminal 5, and controls the process of using the agent for communication with the interlocutor. The agent generation processing unit 15 includes an output destination specification reception unit 16, an agent generation unit 17, an agent output unit 18, and a dialogue processing unit 19. The output destination specification reception unit 16 has a function as an output destination reception means. The output destination specification reception unit 16 receives, from the setting terminal 4, the specification of the output terminal 5 which is the output destination of the agent to be generated. The agent generation unit 17 has a function as an agent generation means. The agent generation unit 17 generates an agent to which a plurality of element information included in the setting file corresponding to the specified agent stored in the setting file storage unit 25, the AI engine, and the speech recognition engine are applied.

[0023] The agent output unit 18 has a function as an agent output means. The agent output unit 18 outputs the agent generated by the agent generation unit 17 to the specified output terminal 5 received by the output destination specification reception unit 16. The dialogue processing unit 19 has functions as a dialogue reception means and a dialogue output means. The dialogue processing unit 19 receives dialogue input data via the output terminal 5. Further, the dialogue processing unit 19 outputs the dialogue output data generated by applying the dialogue input data to the AI engine and the speech recognition engine, in association with the agent output by the agent output unit 18, to the output terminal 5. The communication interface unit 39 is an interface for performing communication with, for example, the setting terminal 4, the output terminal 5, etc. via the communication network N.

[0024] <Setting terminal 4> The setting terminal 4 shown in FIG. 1 is, for example, a personal computer (PC) used by a setter who sets an agent. The setting terminal 4 may also be a portable terminal having the functions of a computer typified by a tablet or a smartphone. Here, in FIG. 1, only one setting terminal 4 is shown, but a plurality of setting terminals 4 can be connected to the agent processing server 1. The setting terminal 4 communicates with the agent processing server 1 via, for example, a web browser. The setting terminal 4 displays, for example, the web screen transmitted by the agent processing server 1, receives the data input and / or selected by the agent setter via the web screen, and transmits it to the agent processing server 1. Although not shown in the figure, the setting terminal 4 includes each part provided in a general computer such as a control unit, a storage unit, an input unit, a display unit, and a communication interface unit.

[0025] <Output terminal 5> The output terminal 5 shown in FIG. 1 is a terminal having a display device used by an interlocutor. In FIG. 1, a PC 5a and a signage 5b are exemplified as the output terminal 5. The PC 5a may be a general personal computer, or may be a portable terminal having the functions of a computer typified by a tablet or a smartphone. The signage 5b is, for example, a large screen device set in the city. In FIG. 1, one PC 5a and one signage 5b are shown as the output terminal 5, but other types of output terminals 5 may exist, and the number thereof may also be plural.

[0026] The output terminal 5 communicates with the agent processing server 1 via, for example, a web browser. The output terminal 5 outputs, for example, the voice data transmitted by the agent processing server 1 via a speaker and outputs the web screen to the display unit. Further, the output terminal 5 transmits the voice data of the speech by the interlocutor acquired via the microphone and the data input via the input unit to the agent processing server 1. Although not shown in the figure, the output terminal 5 includes a microphone and a speaker in addition to each part provided in a general computer such as a control unit, a storage unit, an input unit, a display unit, and a communication interface unit. Note that the input unit and the display unit may be a touch panel display in which their functions are integrated.

[0027] Here, a computer refers to an information processing device equipped with a control unit, a storage device, etc. The agent processing server 1, the setting terminal 4, and the output terminal 5 are all information processing devices equipped with a control unit, a storage unit, etc., and are included in the concept of a computer.

[0028] <Processing of Agent Generation System 100> Next, the processing in the agent generation system 100 will be described. FIG. 3 is a flowchart showing the agent setting process of the agent processing server 1 according to the present embodiment. FIGS. 4 to 6 are diagrams showing display examples on the setting terminal 4 according to the present embodiment. The agent setting process shown in FIG. 3 is performed by the agent setter connecting the setting terminal 4 to the web page provided by the agent processing server 1.

[0029] In step S (hereinafter simply referred to as "S") 11 of FIG. 3, the agent setting unit 11 of the agent processing server 1 outputs a setting screen to the setting terminal 4. An example of the setting screen output to the setting terminal 4 is shown in FIG. 4. The setting screen 40 shown in FIG. 4 is a screen for generating one agent and specifying an output destination. The setting screen 40 is composed of a menu area 41 and a main area 42. The menu area 41 is an area for outputting menus. In the setting screen 40 of FIG. 4, the appearance setting 41a is selected in the menu area 41.

[0030] The main area 42 is an area for outputting setting items (elements) for the selected menu in the menu area 41. In the setting screen 40 of FIG. 4, the main area 42 has an appearance element area 51 and a preview area 52. The appearance element area 51 includes setting buttons 51a to 51c and an apply button 51d. The setting button 51a is related to avatar settings, the setting button 51b is related to scene settings, and the setting button 51c is related to motion settings. The apply button 51d is for finalizing the set content and applying it to the agent. The preview area 52 is an area for previewing the appearance set in the appearance element area 51. In the preview frame 52a of the preview area 52, the appearance of the agent set in the appearance element area 51 is previewed and displayed.

[0031] In S12 of FIG. 3, the agent setting unit 11 determines whether an element item has been received. In the example of the setting screen 40 in FIG. 4 described above, for example, when any one of the setting buttons 51a to 51c is selected by the setter, the agent setting unit 11 determines that an element item has been received. For example, when the setting button 51a is selected, the control unit 10 determines that the element item of the avatar has been received. When an element item has been received (S12: YES), the agent setting unit 11 moves the process to S13. On the other hand, when an element item has not been received (S12: NO), the agent setting unit 11 moves the process to S16. In S13, the element candidate output unit 12 outputs element candidates for the received element item to the setting terminal 4.

[0032] In the setting screen 40 of FIG. 4, when the setting button 51a is selected by the setter, the element candidate output unit 12 of the agent processing server 1 outputs an avatar designation screen 55 shown in FIG. 5 to the setting terminal 4. The avatar designation screen 55 may be, for example, a separate screen output superimposed on the setting screen 40. The avatar specification screen 55 includes a selection area 55a and selection buttons 55b to 55d. In the selection area 55a, a plurality of images (element information) stored in the element storage unit 23, where the element is an avatar (appearance), are output as element candidates. The avatar may be 2D or 3D. Also, the avatar may be a non-human animal or the like. The setter can determine the avatar of the agent by selecting a preferred image from the selection area 55a. Also, when there is no preferred image in the selection area 55a, the setter may operate the selection button 55b to load an image file owned by the setter and use it as the avatar of the agent.

[0033] As another method, when the selection button 55b is operated to load a photo, the control unit 10 of the agent processing server 1 may use an avatar generated from the loaded photo. Further, when the selection button 55b is operated to load text describing the appearance of the agent, the control unit 10 of the agent processing server 1 may use a 3D model automatically generated from the text using a generation AI. The process of generating an avatar from these photos and text may be, for example, that the agent processing server 1 transmits data to an external server that performs the corresponding process, and receives the avatar generated by the external server. The selection button 55c is a button for canceling the specification of the avatar specification screen 55. The selection button 55d is a button for determining the selected or loaded image as element information.

[0034] On the setting screen 40 of FIG. 4, when the setter selects and operates the setting button 51b, a screen for specifying a scene (not shown) is output, and element information where the element is a scene can be determined in the same manner as in the case of the avatar specification screen 55. In the element storage unit 23, a plurality of images of patterns of wallpapers to be set as the background of the avatar, for example, are registered as element information where the element is a scene. Also, on the setting screen 40, when the setter selects and operates the setting button 51c, a screen (not shown) for specifying a motion is output, and element information whose elements are motions can be determined in the same manner as in the case of the avatar specification screen 55. In the element storage unit 23, as element information whose elements are motions, for example, a plurality of images such as a standby motion, a bowing motion, a waving motion, and an explanatory motion, which are examples of the movements of the avatar, are registered.

[0035] For example, on the avatar specification screen 55 of FIG. 5 described above, when the selection button 55d is selected and operated by the setter, in S14 of FIG. 3, the element information reception unit 13 receives the element information. In S15, the agent setting unit 11 updates the setting screen. Specifically, on the setting screen 40 shown in FIG. 4, a character string corresponding to the selected element information is output to the part where the description of the corresponding element item being unset is made. Also, a preview corresponding to the determined element information is displayed in the preview frame 52a. In S16, the agent setting unit 11 determines whether element information has been received for all elements. If element information has been received for all elements (S16: YES), the agent setting unit 11 moves the process to S17. On the other hand, if element information has not been received for all elements (S16: NO), the agent setting unit 11 moves the process to S12 and performs processing on other elements for which element information has not been received. Note that the agent setting unit 11 only needs to receive element information regarding at least appearance, voice color, personality, and knowledge, and in that case, it may proceed to the YES process assuming that element information has been received for all elements. For example, on the setting screen 40 shown in FIG. 4, in addition to the appearance setting 41a in the menu area 41, it is necessary to select and set the inner surface setting 41b.

[0036] FIG. 6 shows an example of the setting screen 40 in which the inner surface setting 41b is selected in the menu area 41. In the setting screen 40 shown in FIG. 6, the main area 42 has an inner surface element area 61. The inner element area 61 includes selection boxes 61a to 61c and an apply button 61d. The selection box 61a is related to the setting of voice and color, the selection box 61b is related to the setting of personality and character, and the selection box 61c is related to role (knowledge). The apply button 61d is for finalizing the setting content and applying it to the agent.

[0037] Note that the selection boxes 61a to 61c of the setting screen 40 illustrated in FIG. 6 are just examples, and element information may be determined by other methods. For example, in the case of voice and color, a file generated using synthetic voice AI based on sample voice data may be specified, or voice may be generated by specifying the parameters of the four elements of sound: propagation, amplitude, frequency, and timbre, and then used. Also, for example, in the case of personality and character, a generation AI may be used, and text describing the personality and character may be input to make settings according to the content of the text. Also, for example, in the case of a role, a learning model learned by inputting a document corresponding to the role into a generation AI may be specified. These processes may be, for example, the agent processing server 1 sending data to an external server that performs the corresponding process and receiving the data and files generated by the external server.

[0038] In S17 of FIG. 3, the control unit 10 (setting file creation unit 14) creates a setting file. For example, the control unit 10 newly creates a setting file in the setting file storage unit 25 and assigns a file ID. Then, the control unit 10 stores the received element information in the created setting file. Also, the control unit 10 registers the file ID in association with the setting ID of the corresponding settler in the settler information storage unit 24. By doing so, the element information of each element constituting one agent is registered in one setting file. And the setting file can be associated with the settler. After that, the control unit 10 ends this process.

[0039] As described above, the agent processing server 1 outputs the setting screen 40 to the setting terminal 4, and the setter can set the agent only by performing an operation of selecting the element information to be applied to the agent from the plurality of element information of each element. Therefore, the setter can set the agent more easily by the agent processing server 1. In addition, since the setter designates the element information for a plurality of element items regarding the appearance and the inner surface, the agent can be set in more detail.

[0040] Next, the process of generating and outputting the agent using the setting file created in the above process will be described. FIG. 7 is a flowchart showing the agent output process of the agent processing server 1 according to the present embodiment. FIG. 8 is a diagram showing a display example on the setting terminal 4 according to the present embodiment. FIG. 9 is a diagram showing a specific example on the output terminal 5 according to the present embodiment. The agent output process shown in FIG. 7 may be performed following the agent setting process described in FIG. 3, or may be performed separately.

[0041] In S21 of FIG. 7, the agent generation processing unit 15 of the agent processing server 1 receives the selection of the setting file from the setting terminal 4. After connecting the setting terminal 4 to the agent processing server 1, when the setter inputs the setter ID, for example, although not shown, a screen for selecting one setting file from the setting files associated with the setter ID is output. Then, by the selection operation of one setting file by the setter, the agent generation processing unit 15 receives the selection of the setting file. In S22, the agent generation processing unit 15 outputs the setting screen to the setting terminal 4.

[0042] A display example of the setting screen on the setting terminal 4 is shown in FIG. 8. The setting screen 40 shown in FIG. 8 is a screen for specifying the output terminal 5 which is the output destination of the agent. In the menu area 41 of the setting screen 40, the distribution destination setting 41c is selected. Also, the main area 42 has a terminal area 71. The terminal area 71 includes selection buttons 71a and 71b and a distribution button 71c. The selection buttons 71a and 71b correspond to the respective output terminals 5. Note that the information of the output terminal 5 corresponding to the selection buttons 71a and 71b to be output to the terminal area 71 is registered in a screen (not shown) in advance. Also, the information of the registered output terminal 5 may be stored in, for example, the setter information storage unit 24. The distribution button 71c is a button for outputting the agent to the selected output terminal 5.

[0043] In S23 of FIG. 7, the output destination designation reception unit 16 receives the selection of the output destination. In the example of the setting screen 40 of FIG. 8 described above, when the distribution button 71c is selected while either one of the selection buttons 71a and 71b is selected, the output destination designation reception unit 16 receives the selection of the output terminal 5 as the output destination. In S24, the agent generation unit 17 generates an agent that applies the plurality of element information stored in the setting file received in S21, the AI engine, and the speech recognition engine. In S25, the agent output unit 18 outputs the generated agent to the output terminal 5 received in the process of S23. In S26, the dialogue processing unit 19 performs dialogue processing. More specifically, when the dialogue input data is received via the output terminal 5, the dialogue processing unit 19 applies the dialogue input data to the AI engine and the speech recognition engine to generate dialogue output data, and outputs the dialogue output data to the output terminal 5 in association with the output agent.

[0044] FIG. 9 shows an example of the display of the signage 5b that transitions by the processing of the dialogue processing unit 19. The screen 81 shown in Fig. 9(A) shows the state in which the agent 81a generated by the agent output unit 18 is output. The screen 81 outputs the agent 81a generated based on a plurality of element information stored in the setting file as an avatar, and the question of the agent 81a is output as a balloon 81b. Note that the dialogue processing unit 19 may output the same content as the balloon 81b in voice. In that case, voice based on element information such as voice tone and personality stored in the setting file is output.

[0045] The screen 82 shown in Fig. 9(B) is when the interlocutor asks a question to the agent 81a. When the interlocutor speaks a question, a new balloon 82a is output, and the dialogue processing unit 19 receives the dialogue input data as voice data from the sign 5b, and outputs the text generated by the speech recognition engine to the balloon 82a. Then, the dialogue processing unit 19 causes the AI engine to execute a process of automatically generating a response text for the text using the element information stored in the setting file, causes the speech recognition engine to generate dialogue output data which is voice data from the response text, and outputs the generated dialogue output data to the sign 5b.

[0046] The screen 83 shown in Fig. 9(C) is when the agent 81a answers a question. The balloon 83a on the screen 83 is an answer to the interlocutor's question. The dialogue processing unit 19 outputs the balloon 83a on the screen 83 and also outputs the dialogue output data as voice. Returning to Fig. 7, then, for example, upon receiving a stop instruction from the output terminal 5 such as power-off of the output terminal 5, the agent generation processing unit 15 ends this process.

[0047] As described above, according to the agent processing server 1 of this embodiment, there are the following effects. (1) Receive element information corresponding to each of a plurality of elements constituting one agent from the setting terminal 4, store the plurality of element information for the received one agent in a setting file, and apply the plurality of element information for the one agent stored in the setting file to an AI engine and a speech recognition engine to generate one agent. Therefore, if the setter designates element information corresponding to each of a plurality of elements constituting one agent, an agent corresponding to the element information can be generated, and a mechanism can be provided that enables the setter to set the agent more easily and in more detail.

[0048] (2) Include an element storage unit 23 that stores a plurality of element information corresponding to each of a plurality of elements, output the plurality of element information for each element stored in the element storage unit 23 to the setting terminal 4, and receive, from the setting terminal 4, the element information selected by the setter of the setting terminal 4 from the plurality of element information for each element. Therefore, since the setter can determine the element information for each element of the agent by performing a selection operation, the detailed setting of the agent can be performed more easily.

[0049] (3) Receive, from the setting terminal 4, the element information corresponding to the element input by the setter in the setting terminal 4. Therefore, in addition to the plurality of element information prepared in advance, the element information of the elements of the agent input by the setter can be incorporated, so that a mechanism can be provided that enables the setter to customize the configuration of the agent.

[0050] (4) The elements include at least appearance, voice color, personality, and knowledge, and the element information includes the setting content of the elements. Therefore, a mechanism can be provided that enables not only the appearance of the agent but also the internal settings to be made.

[0051] (5) Receive, from the setting terminal 4, the designation of an output terminal 5 for outputting the generated one agent, and output the generated one agent to the designated output terminal 5 received. Therefore, it is possible to easily specify the output terminal 5 that outputs the agent from the setting terminal 4. And since the generated agent can be output to the specified output terminal 5, it is convenient.

[0052] (6) The interactive input data is received via the output terminal 5, and the interactive output data generated by applying the interactive input data to the AI engine and the speech recognition engine is output to the output terminal 5 in association with the output agent. And the AI engine executes the process of automatically generating the answer text for the input text of the interactive input data using the element information for one agent stored in the setting file. Also, the speech recognition engine generates text from the received interactive input data and executes the process of generating interactive output data from the answer text generated by the AI engine. Therefore, since the agent output to the output terminal 5 makes the answer to the interactive input data of the interlocutor correspond to the element information, it is possible to make the agent have a natural conversation.

[0053] The embodiments of the present invention have been described above, but the present invention is not limited to the above-described embodiments. Also, the effects described in the embodiments are merely an enumeration of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the embodiments. Incidentally, the above-described embodiments and the deformation modes described later can be used in appropriate combination, but detailed description thereof is omitted.

[0054] (Modification) (1) In this embodiment, the example in which the element information corresponding to each of the plurality of elements constituting the agent is newly stored in the setting file has been described, but the present invention is not limited to this. The agent processing server may have a function of changing at least a part of the plurality of element information stored in the setting file. In that case, the file reception means receives a selection of a configuration file stored in the configuration file storage unit from the configuration terminal, and the element information output means outputs a plurality of pieces of element information stored in the configuration file corresponding to the reception to the configuration terminal. Then, the element information reception means receives the changed element information from the configuration terminal, and the configuration storage means stores the received changed element information in the configuration file. By such processing, a part of the agent can be easily customized to generate a new agent.

[0055] (2) In the present embodiment, the case where the agent processing server 1 includes the AI engine unit 31 and the speech recognition engine unit 32 has been described as an example, but the present invention is not limited to this. The AI engine unit and the speech recognition engine unit may be provided in an external server connected via the communication network N, and the agent processing server may perform processing in cooperation with the external server.

[0056] (3) In the present embodiment, the case where an agent is output to one output terminal has been described as an example, but the present invention is not limited to this. The agent may be output to a plurality of output terminals. By doing so, at each of the plurality of output terminals, the agent can respond to the dialogue of the interlocutor.

Explanation of Reference Numerals

[0057] 1 Agent processing server 4 Configuration terminal 5 Output terminal 5a PC 5b Signage 10 Control unit 11 Agent configuration unit 12 Element candidate output unit 13 Element information reception unit 14 Configuration file creation unit 15 Agent generation processing unit 16 Output destination specification reception unit 17 Agent generation unit 18 Agent output unit 19 Dialogue processing unit 20 Memory unit 21a Agent program 23 Element memory unit 24 Setter information memory unit 25 Configuration file memory unit 31 AI engine unit 32 Speech recognition engine unit 40 Settings screen 100 Agent generation system

Claims

1. Element information reception means for receiving, from a setting terminal, element information corresponding to each of a plurality of elements constituting one agent; Setting storage means for storing, in a setting file, the plurality of pieces of element information for the one agent received by the element information reception means; Agent generation means for generating the one agent to which the plurality of pieces of element information for the one agent stored in the setting file and an AI engine and a voice recognition engine are applied; An agent processing apparatus comprising the above.

2. In the agent processing apparatus according to Claim 1, An element storage unit that stores a plurality of pieces of the element information corresponding to each of the plurality of elements; Element candidate output means for outputting, to the setting terminal, the plurality of pieces of element information for each of the elements stored in the element storage unit; Comprising, The element information reception means receives, from the setting terminal, the element information selected by the user of the setting terminal from the plurality of pieces of element information for each element. An agent processing apparatus.

3. In the agent processing apparatus according to Claim 2, The element information reception means receives, from the setting terminal, the element information corresponding to the element input by the user of the setting terminal in the setting terminal. An agent processing apparatus.

4. In the agent processing apparatus according to any one of Claims 1 to 3, The elements include at least appearance, voice color, personality, and knowledge, and the element information includes the setting content of the elements. An agent processing apparatus.

5. In the agent processing apparatus according to any one of Claims 1 to 3, Output destination reception means for receiving, from the setting terminal, a designation of an output terminal for outputting the one agent to be generated; Agent output means for outputting the one agent generated by the agent generation means to the designated output terminal received by the output destination reception means; Comprising, an agent processing apparatus.

6. In the agent processing apparatus according to Claim 5, Dialog reception means for receiving dialog input data via the output terminal; Dialog output means for outputting, to the output terminal, in association with the agent output by the agent output means, dialog output data generated by applying the dialog input data to the AI engine and the voice recognition engine; Comprising, an agent processing apparatus.

7. In the agent processing apparatus according to claim 6, The AI engine executes a process of automatically generating a response text for the input text of the dialogue input data using the element information for the one agent stored in the setting file. An agent processing apparatus.

8. In the agent processing apparatus according to claim 7, The voice recognition engine generates text from the dialogue input data received by the dialogue receiving means, and executes a process of generating the dialogue output data from the response text generated by the AI engine. An agent processing apparatus.

9. In the agent processing apparatus according to any one of claims 1 to 3, A file reception means for receiving a selection of the setting file stored by the setting storage means from the setting terminal, Element information output means for outputting the plurality of element information stored in the setting file corresponding to the reception by the file reception means to the setting terminal, Comprising, The element information reception means receives the changed element information from the setting terminal, The setting storage means stores the changed element information received by the element information reception means in the setting file. An agent processing apparatus.

10. An agent processing apparatus according to any one of claims 1 to 3, A setting terminal communicably connected to the agent processing apparatus, An agent generation system comprising.

11. A computer, Element information reception means for receiving element information corresponding to each of a plurality of elements constituting one agent from a setting terminal, Setting storage means for storing the plurality of element information for the one agent received by the element information reception means in a setting file, Agent generation means for generating the one agent to which the plurality of element information for the one agent stored in the setting file, an AI engine, and a voice recognition engine are applied, A program for functioning as.

Images