Device and method

Abstract

A device according to the present invention comprises: an acquisition unit that acquires information pertaining to an existence area in which an autonomous agent acting autonomously in a virtual space exists, among a plurality of areas included in the virtual space, and information pertaining to an object which exists in the existence area; a determination unit that determines an action of the autonomous agent on the basis of the information pertaining to the existence area and the information pertaining to the object which exits in the existence area; and a prompt generation unit that generates a prompt for generating information pertaining to the action of the autonomous agent on the basis of the information acquired by the acquisition unit.

Description

Device and Method

[0001] The present invention relates to a device and a method for determining the behavior of an autonomous agent that autonomously acts within a virtual space.

[0002] Patent Document 1 describes a device that operates a non-player character to move within a game space toward a movement target point associated with the current in-game time, and causes an enemy character that has reached the movement target point to execute an operation associated with the movement target point.

[0003] Japanese Patent Application Laid-Open No. 2012-213587

[0004] The technique described in Patent Document 1 causes a non-player character, which is an autonomous agent, to perform a specific operation associated with a movement target point. Therefore, the behavior of the autonomous agent becomes extremely limited.

[0005] Therefore, an object of the present disclosure is to provide a device and a method that can cause an autonomous agent to perform various operations.

[0006] The device of the present disclosure includes an acquisition unit that acquires information regarding an existence area in which an autonomous agent that autonomously acts in the virtual space exists among a plurality of areas included in the virtual space, and information regarding an object existing in the existence area, a determination unit that determines the behavior of the autonomous agent based on the information regarding the existence area and the information regarding the object existing in the existence area, and a prompt generation unit that generates a prompt for generating information regarding the behavior of the autonomous agent based on the information acquired by the acquisition unit.

[0007] According to the present disclosure, an autonomous agent can be caused to perform various operations.

[0008] Figure 1 is a diagram showing the system configuration of a game system including an autonomous agent action decision device according to an embodiment of the present disclosure. Figure 2 is a block diagram showing the functional configuration of the autonomous agent action decision device. Figure 3 is a diagram showing an example of information stored in the area database. Figure 4 is a diagram showing an example of information stored in the object database. Figure 5 is a diagram showing an example of information stored in the personality database. Figure 6 is a diagram showing an example of information stored in the action database. Figure 7 is a sequence diagram showing an example of a processing flow. Figure 8 is a diagram showing an example of a prompt. Figure 9 is a diagram showing an example of the output of a large-scale language model. Figure 10 is a diagram showing an example of the hardware configuration of a test data generation device according to an embodiment of the present disclosure.

[0009] Embodiments of this disclosure will be described with reference to the attached drawings. Where possible, the same parts will be denoted by the same reference numerals, and redundant descriptions will be omitted.

[0010] Figure 1 shows the system configuration of a game including an autonomous agent action decision device 100 according to an embodiment of the present disclosure. As shown in the figure, the system comprises the autonomous agent action decision device 100, a large-scale language model 200, a game server 300, an area database 400, an object database 500, a personality database 600, an action database 700, and a data parser 800. Each of these servers and databases is configured to communicate with each other via a network.

[0011] The game server 300 operates in a virtual space with multiple areas, such as the metaverse, where it controls agents such as avatars operated by users, and autonomous agents, which are non-player characters that act autonomously.

[0012] The user operates the user terminal 900 to make an agent act in the virtual space provided by the game server 300. The user terminal 900 transmits information about its operations to the game server 300. Based on the operation information received from the user terminal 900, the game server 300 makes the agent act in the virtual space.

[0013] The game server 300 transmits information regarding the autonomous agent's location in the virtual space to the autonomous agent action decision device 100. Upon receiving the information regarding the autonomous agent's location from the game server 300, the autonomous agent action decision device 100 obtains information regarding the area where the autonomous agent resides, information regarding objects present in that area, information regarding the autonomous agent's personality, and information regarding actions the autonomous agent can take in that area, from among multiple areas included in the virtual space. Information regarding the area is obtained, for example, from the area database 400. Information regarding objects present in the area is obtained, for example, from the object database 500. Information regarding the autonomous agent's personality is obtained, for example, from the personality database 600. Information regarding actions the autonomous agent can take in that area is obtained, for example, from the action database 700.

[0014] The autonomous agent action decision device 100 determines the autonomous agent's actions based on acquired information about the presence area, information about objects present in the presence area, information about the autonomous agent's personality, and information about actions the autonomous agent can take in the presence area.

[0015] The autonomous agent action decision device 100 generates a prompt to instruct the generating AI to generate information about the autonomous agent's actions, based on acquired information about the presence area, information about objects present in the presence area, information about the autonomous agent's personality, and information about actions the autonomous agent can take in the presence area, in order to determine the autonomous agent's actions. The autonomous agent action decision device 100 transmits the generated prompt to the large-scale language model 200. The large-scale language model 200 generates information about the autonomous agent's actions based on the prompt received from the autonomous agent action decision device 100, and transmits the generated information about the autonomous agent's actions to the autonomous agent action decision device 100. The autonomous agent action decision device 100 determines the autonomous agent's actions based on the information about the autonomous agent's actions received from the large-scale language model 200.

[0016] When the autonomous agent action decision device 100 decides on an action for the autonomous agent, it transmits information about the decided autonomous agent action to the data parser 800. Based on the information about the autonomous agent action received from the autonomous agent action decision device 100, the data parser 800 creates request information for the game server 300 and transmits the created request information to the autonomous agent action decision device 100. The autonomous agent action decision device 100 transmits the request information received from the data parser 800 to the game server 300. Based on the request information received from the autonomous agent action decision device 100, the game server 300 causes the autonomous agent to act in the virtual space.

[0017] The large-scale language model 200 is a generative AI model that, in response to input from a prompt containing input information, generates content according to the instructions, context, questions, and output format indicated by the prompt, and returns that content as response information. In this disclosure, a prompt is information indicating instructions or questions entered by a user in an interactive system such as a dialogue with a generative AI model or a command-line interface (CLI). The prompt may also include input information, in which case the generative AI model generates response information targeting the input information. The generative AI model may be an interactive AI model that includes, for example, a large-scale language model (LLM) and a user interface (UI) for dialogue with the user, enabling text chat or voice chat with the user. Examples of such generative AI models include ChatGPT, GPT®-3.5, GPT-4V, GPT-4o, PaLM2, and tsuzumi.

[0018] Furthermore, all or some of the large-scale language model 200, game server 300, area database 400, object database 500, personality database 600, behavior database 700, and data parser 800 may be located in the autonomous agent behavior decision device 100 or the user terminal 900. Alternatively, the user terminal 900 may have the functionality of the autonomous agent behavior decision device 100 and function as such. Some generated AI models, like Tsuzumi, have the generated AI model located within the user terminal 900. In this type, the RAG (Retrieval-Augmented Generation) application is also provided on the user terminal 900. However, the information (database) accessed by the RAG application may be located within the user terminal 900 or on the network. There are also types, like ChatGPT, where the generated AI model is located on the network. In this type, the RAG application is provided on the user terminal 900. However, the information (database) that RAG accesses resides on the network.

[0019] Figure 2 is a block diagram showing the functional configuration of the autonomous agent action decision device 100. As shown in Figure 2, the autonomous agent action decision device 100 is composed of a reception unit 101, an acquisition unit 102, a decision unit 103, a prompt generation unit 104, a transmission / reception unit 105, and a request unit 106.

[0020] The reception unit 101 receives information from the game server 300 that identifies the autonomous agent, as well as information regarding the autonomous agent's location within the virtual space. The virtual space contains multiple areas, such as a central plaza, a fishing area, and a shopping area.

[0021] The acquisition unit 102 acquires information about the presence area, information about objects present in the presence area, information about the personality of the autonomous agent, and information about the actions the autonomous agent can take.

[0022] The acquisition unit 102 acquires information about the area where the autonomous agent is located, for example, based on information about the autonomous agent's location and information stored in the area database 400.

[0023] The area database 400 stores information about each of several areas. This area information includes, for example, information indicating what the area is for. Figure 3 shows an example of the information stored in the area database 400. In the example shown in Figure 3, the area database 400 stores information about three areas. In the example shown in Figure 3, the information stored in the area database 400 includes the area identifier (area_id), the area name (name), the location coordinates of the area's representative point (location), and the area description (description).

[0024] The area identifier (area_id) field stores information to identify each of multiple areas. In the example shown in Figure 3, the area identifier field stores "central_square_01" as the identifier for the first area, "fishing_area_01" as the identifier for the second area, and "shopping_area_01" as the identifier for the third area.

[0025] The "Area Name" field stores the names of multiple areas. In the example shown in Figure 3, the "Area Name" field stores "central square" as the name of the first area, "fishing area" as the name of the second area, and "shopping area" as the name of the third area.

[0026] The "location" field for the representative point of an area stores information about the location of each representative point of multiple areas. In the example shown in Figure 3, the "location" field for the representative point of an area stores "0.2,0.4,0.6" as the location coordinates of the first area's representative point, "-50.2,20.4,30.6" as the location coordinates of the second area's representative point, and "100.3,267.1,534.3" as the location coordinates of the third area's representative point.

[0027] The "Description" field stores the descriptions for each of the multiple areas. In the example shown in Figure 3, the "Description" field stores the coordinates of the representative point for the first area as "A place where people gather and talk," the coordinates of the representative point for the second area as "People can enjoy fishing here," and the coordinates of the representative point for the third area as "A place where player_avatar enjoy shopping."

[0028] The acquisition unit 102 then acquires, for example, information about the area corresponding to the location of the autonomous agent acquired by the reception unit 101 from the information stored in the area database 400, as information about existing areas. The information about existing areas acquired by the acquisition unit 102 can be at least one of the above-mentioned area identifier (area_id), area name (name), and area description (description).

[0029] The acquisition unit 102 acquires information about objects present in the presence area, for example, based on information about the autonomous agent's location and information stored in the object database 500.

[0030] The object database 500 stores information about objects present in each of multiple areas. This information about objects present in each of multiple areas includes, for example, information indicating the actions that agents and autonomous agents can take towards those objects. Figure 4 shows an example of the information stored in the object database 500. In the example shown in Figure 4, the object database 500 stores information about five objects. In the example shown in Figure 4, the information stored in the object database 500 includes the object identifier (object_id), the object name (name), the object's location coordinates (location), and the actions that can be taken towards the object (affordance).

[0031] The object identifier (object_id) field stores information to identify the object. In the example shown in Figure 4, the object identifier field stores "chair_01" as the identifier for the first object, "advulertisement_01" as the identifier for the second object, "player_avulater_01" as the identifier for the third area, "AI_avulater_01" as the identifier for the fourth area, and "fwish_spot_01" as the identifier for the fifth area.

[0032] The "name" field stores the names of the objects. In the example shown in Figure 4, the "name" field stores the names of the first object as "chair", the second object as "megumi", the third object as "player avatar", the fourth object as "warp_zone", and the fifth object as "fish_spot".

[0033] The "location" field stores information about the object's position. In the example shown in Figure 4, the "location" field stores the following coordinates for each object: "0.2,0.4,0.6" for the first object, "0.2,0.4,0.6" for the second object, "0.2,0.4,0.6" for the third object, "0.2,0.4,0.6" for the fourth object, and "0.2,0.4,0.6" for the fifth object.

[0034] The "affordance" item for an object stores the actions that agents and autonomous agents can take on that object. In the example shown in Figure 4, the "affordance" item for an object stores the following actions for the first object: "The avatar can sit down by using the “sit_on_something” action with the object_id specified.", "The avatar can see advertisement by using the “see_advertisement” action with the object_id specified.", "The avatar can speak to them by using the “chat_avatar” action with the object_id specified.", "The avatar can speak to them by using the “chat_avatar” action with the object_id specified.", and "The avatar can fish by using the “fish” action with the object_id specified."

[0035] The acquisition unit 102 then acquires, for example, information from the object database 500 that corresponds to the location of the autonomous agent acquired by the reception unit 101, as information about objects existing in the existence area. The information about objects existing in the existence area acquired by the acquisition unit 102 can be at least one of the above-mentioned object identifier (area_id), object name (name), and action that can be taken towards the object (affordance).

[0036] The acquisition unit 102 acquires information about the personality of an autonomous agent, for example, based on information that identifies the autonomous agent and information stored in the personality database 600.

[0037] The personality database 600 stores information about the personality of autonomous agents. This information includes, for example, the autonomous agent's role, behavioral guidelines, and information that the autonomous agent should know. Figure 5 shows an example of the information stored in the personality database 600. In the example shown in Figure 5, the personality database 600 stores information about the personalities of two autonomous agents. In the example shown in Figure 5, the information stored in the personality database 600 includes the autonomous agent's identifier (avatar_id), name, role, behavioral guidelines, and knowledge.

[0038] The avatar_id field stores information to identify the autonomous agent. In the example shown in Figure 5, the avatar_id field stores "sales_clerk_01" as the identifier for the first autonomous agent and "live_participant_01" as the identifier for the second autonomous agent.

[0039] The "Name of Autonomous Agent" field stores the names of the autonomous agents. In the example shown in Figure 5, the "Name of Autonomous Agent" field stores "satoru" as the name of the first autonomous agent and "megumi" as the name of the second autonomous agent.

[0040] The "role" field for autonomous agents stores information about the role that autonomous agents play within the virtual space. In the example shown in Figure 5, the "role" field for autonomous agents stores "You are a sales clerk at a store. My job is to sell products to users." as the role of the first autonomous agent, and "You are a live participant." as the role of the second autonomous agent.

[0041] The "Affordance" section of the autonomous agent's behavioral guidelines stores the guidelines for the autonomous agent's actions within the virtual space. In the example shown in Figure 5, the first autonomous agent's behavioral guidelines are stored as "Please act within {area name}. If you find a player_avatar, please actively talk to them. However, please don't talk to the same user over and over again." and the second autonomous agent's behavioral guidelines are stored as "Please use a lot of emotes to liven up the live performance. Please refrain from talking to others as it will disturb the live performance."

[0042] The items of information (knowledge) that the autonomous agent should know store information that the autonomous agent should know, such as information related to the autonomous agent. In the example shown in FIG. 5, in the item of information that the autonomous agent should know, "The products you are selling are: (1)Bread (2)Sushi (3)Sandw" is the information that the first autonomous agent should know, and "This concert is by the band "BBB" and is titled "BBB 1st Anniversary". Member Araki is absent." is the information that the second autonomous agent should know, and they are each stored.

[0043] Then, the acquisition unit 102 acquires, as information related to the personality of the autonomous agent, information related to the autonomous agent corresponding to the information for identifying the autonomous agent acquired by the reception unit 101, among the information stored in the personality database 600, for example. As the information related to the configuration of the autonomous agent acquired by the acquisition unit 102, at least one of the above-described identifier (area_id) of the autonomous agent, name (name) of the autonomous agent, role (role) of the autonomous agent, behavioral guidelines (behavioral guidelines) of the autonomous agent, and information (knowledge) that the autonomous agent should know can be used.

[0044] The acquisition unit 102 acquires information related to actions that the autonomous agent can take, based on the information stored in the action database 700, for example.

[0045] The behavior database 700 stores information about the actions that the autonomous agent can take in each of the multiple areas. This information about the actions that the autonomous agent can take in each of the multiple areas includes, for example, information that indicates specific actions that the autonomous agent can take. Figure 6 shows an example of the information stored in the behavior database 700. In the example shown in Figure 6, the behavior database 700 stores information about seven actions. In the example shown in Figure 6, the information stored in the behavior database 700 includes the action identifier (action_id), the action type (classification), the action name (name), and the data in JSON (JavaScript® Object Notation) format (json_format).

[0046] The action_id field stores information to identify an action. In the example shown in Figure 6, the action_id field stores "01" as the identifier for the first action, "02" as the identifier for the second action, "03" as the identifier for the third action, "04" as the identifier for the fourth action, "05" as the identifier for the fifth action, "06" as the identifier for the sixth action, and "07" as the identifier for the seventh action.

[0047] The behavior type (classification) item stores whether the behavior is a smallest unit of behavior (atom) or a behavior composed of a combination of atoms (molecure). In the example shown in Figure 6, the behavior type item stores "atom" as the first behavior type, "atom" as the second behavior type, "atom" as the third behavior type, "atom" as the fourth behavior type, "atom" as the fifth behavior type, "molecure" as the sixth behavior type, and "molecure" as the seventh behavior type.

[0048] In the item of the action name (action_name), the name of the action is stored. In the example shown in FIG. 6, in the item of the action name, "chat" is stored as the name of the first action, "move_to" is stored as the name of the second action, "move_around" is stored as the name of the third action, "sit_down" is stored as the name of the fourth action, "emote" is stored as the name of the fifth action, "chat_avatar" is stored as the name of the sixth action, and "sit_on_something" is stored as the name of the seventh action, respectively.

[0049] In the item of the JSON format data (json_format), the JSON format data indicating the action is stored. In the example shown in FIG. 6, in the item of the JSON format data indicating the action, "{action: chat}" is stored as the JSON format data indicating the first action, "{action: move_to, object_id: object_id}" is stored as the JSON format data indicating the second action, "{action: move_around, object_id: None}" is stored as the JSON format data indicating the third action, "{action: sit_down, object_id}" is stored as the JSON format data indicating the fourth action, "{action: emote, emote_id: emote_id}" is stored as the JSON format data indicating the fifth action, "{action: move_to, object_id: object_id}, {action: chat}" is stored as the JSON format data indicating the sixth action, and "{action: move_to, object_id: object_id}, {action: sit_down, object_id: object_id}" is stored as the JSON format data indicating the seventh action, respectively.

[0050] The acquisition unit 102 then acquires information on at least one action from the information stored in the action database 700 as information on actions that the autonomous agent can take. The information on actions that the autonomous agent can take, acquired by the acquisition unit 102, can be at least one of the following: the action identifier (action_id), the action type (classification), the action name (name), and data in JSON (JavaScript Object Notation) format (json_format).

[0051] The decision unit 103 determines the autonomous agent's actions based on the information about the existence area acquired by the acquisition unit 102, the information about objects present in the existence area, the information about the autonomous agent's personality, and the information about actions the autonomous agent can take. For example, the decision unit 103 determines the autonomous agent's actions based on the information about the autonomous agent's actions generated by the large-scale language model 200.

[0052] The prompt generation unit 104 generates prompts for generating information about the autonomous agent's actions based on the information acquired by the acquisition unit 102. The information acquired by the acquisition unit 102 is, as described above, information about the presence area, information about objects present in the presence area, information about the autonomous agent's personality, and information about actions the autonomous agent can take. The prompts generated by the prompt generation unit 104 may be based on information about the presence area and information about objects present in the presence area, or they may be based on information that includes at least one of the information about the autonomous agent's personality and information about actions the autonomous agent can take.

[0053] The prompt includes information about the area of ​​existence and information about objects present in that area. The prompt may also include at least one of information about the personality of the autonomous agent and information about actions the autonomous agent can take. Furthermore, the prompt may include at least one of an instruction for the large-scale language model 200 to generate an action for the autonomous agent, and the output format of the action for the autonomous agent generated by the large-scale language model 200.

[0054] The transmitting / receiving unit 105 sends the generated prompt to the large-scale language model 200 and obtains information from the large-scale language model 200 regarding the behavior of the autonomous agent generated by the large-scale language model 200.

[0055] The decision unit 103 determines the autonomous agent's actions based on the information about the autonomous agent's actions acquired by the transmission / reception unit 105.

[0056] The request unit 106 transmits information about the autonomous agent's actions determined by the decision unit 103 to the data parser 800. The request unit 106 retrieves JSON format data corresponding to the autonomous agent's actions determined by the decision unit 103 from the action database 700 and transmits this retrieved JSON format data to the data parser 800 as information about the autonomous agent's actions. The data parser 800 converts the information about the autonomous agent's actions received from the request unit 106 into a data structure usable by the game server 300 to create request information and transmits the created request information to the request unit 106. The request unit 106 transmits the request information received from the data parser 800 to the game server 300. Based on the request information received from the request unit 106, the game server 300 causes the autonomous agent to act in the virtual space.

[0057] Next, we will explain the operation of the autonomous agent action decision device 100 configured in this way, that is, the method for determining the autonomous agent's action. Figure 7 is a sequence diagram showing an example of the processing flow.

[0058] As shown in Figure 7, the reception unit 101 receives information from the game server 300 that identifies the autonomous agent and information that identifies the autonomous agent's location in the virtual space (S101).

[0059] Next, the acquisition unit 102 accesses the area database 400 (S102). The acquisition unit 102 then retrieves from the area database 400 information that corresponds to the location of the autonomous agent acquired by the reception unit 101, as information about objects present in the area (S103).

[0060] Furthermore, the acquisition unit 102 accesses the object database 500 (S104). The acquisition unit 102 then obtains information from the object database 500 regarding the object corresponding to the location of the autonomous agent acquired by the reception unit 101, from among the information stored in the object database 500, as information regarding the object present in the presence area (S105).

[0061] Furthermore, the acquisition unit 102 accesses the personality database 600 (S106). The acquisition unit 102 then obtains from the personality database 600 information that corresponds to the information used to identify the autonomous agent acquired by the reception unit 101, as information regarding the personality of the autonomous agent (S107).

[0062] Furthermore, the acquisition unit 102 accesses the behavior database 700 (S108). The acquisition unit 102 then acquires information on at least one behavior from the information stored in the behavior database 700 as information on actions that the autonomous agent can take (S109).

[0063] Furthermore, the acquisition of information regarding the presence area, information regarding objects present in the presence area, information regarding the personality of the autonomous agent, and information regarding the actions that the autonomous agent can take may be carried out in any order, and some or all of this information may be acquired simultaneously.

[0064] Next, the prompt generation unit 104 generates a prompt for generating information about the autonomous agent's actions based on the information acquired by the acquisition unit 102 (S110).

[0065] Next, the transmitting / receiving unit 105 transmits the prompt generated by the prompt generation unit 104 to the large-scale language model 200 (S111). Then, the transmitting / receiving unit 105 obtains information from the large-scale language model 200 regarding the behavior of the autonomous agent generated by the large-scale language model 200 (S112).

[0066] Next, the decision unit 103 determines the autonomous agent's action based on the information about the autonomous agent's actions acquired by the transmission / reception unit 105 (S113).

[0067] Next, the request unit 106 accesses the behavior database 700 (S114). Then, the request unit 106 obtains JSON format data corresponding to the determined autonomous agent's behavior from the behavior database 700 (S115).

[0068] Next, the request unit 106 sends the acquired JSON format data to the data parser 800 (S116). Then, the request unit 106 obtains request information from the data parser 800 that has been converted into a data structure usable by the game server 300 (S117).

[0069] Next, the request unit 106 sends the request information obtained from the data parser 800 to the game server 300 as information regarding the actions of the autonomous agent (S118).

[0070] Next, the game server 300 causes the autonomous agent to act in the virtual space based on the request information received from the request unit 106.

[0071] Here, we will describe an example of a prompt generated by the prompt generation unit 104 of the autonomous agent action decision device 100. Figure 8 shows an example of a prompt. The prompt shown in Figure 8 includes information on decision-making instructions ("Instruction"), personality ("Personality"), perception and cognition ("Awareness"), and output format ("Output").

[0072] The decision instruction in this prompt, "Instruction," is a set of instructions that directs the large-scale language model 200 to determine the actions of the autonomous agent. This decision instruction includes the role and the task.

[0073] The role of this prompt includes information about the role of the large-scale language model 200. In the section for the role of this prompt, the information about the role of the large-scale language model 200 is listed as "You are the one who decides the avatar's actions."

[0074] This prompt's task contains information about the tasks of the large-scale language model 200. The task column of this prompt lists the following tasks for the large-scale language model 200: "Decide the action according to the avatar's surrounding objects, state, and space described in cognition," "Decide the action according to the role, action guidelines, and memory described in personality," and "Generate appropriate action instructions in text based on the action guidelines."

[0075] The "Personality" field in this prompt contains information about the autonomous agent's personality. This includes information about the autonomous agent's personality, specifically its role and guidelines for action.

[0076] The "role" field in this prompt contains information about the autonomous agent's role. The prompt's role field includes the following information about the autonomous agent's role: "• You are a salesperson in a store. My job is to sell products to users."

[0077] The "Guidelines for Action" section of this prompt contains information regarding the autonomous agent's behavior. The section for the autonomous agent's behavior includes the following: "Please act within shopping_area_01. If you find a player_abatar, please actively talk to them. However, please don't talk to the same user repeatedly."

[0078] The "Awareness" of this prompt includes "Your state," "Current location," "Objects," and "Selectable actions."

[0079] The "Your state" field in this prompt contains information about the autonomous agent's state. In this prompt's state field, the autonomous agent's state is listed as "idle: Do nothing".

[0080] The "Current location" field in this prompt contains information about the location where the autonomous agent resides. The field for this prompt's location contains the information "shopping_area_01: The place where the avatar enjoys shopping".

[0081] The "Objects" object in this prompt contains information about the objects present in the presence area. The object field in this prompt lists the following information about the objects present in the presence area: "Advertisement: {object_id:advertisement_1, affodance:The avatar can see advertisement by using the “see_advertisement” action with the object_id specified.}", "Chair: {object_id:chair_1, affodance:The avatar can sit down by using the “sit_on_something” action with the object_id specified.}", "Player: {object_id:player_avatar_1, affodance:The avatar can speak to them by using the “chat_avatar” action with the object_id specified.}", and "Player:{object_id:player_avatar_2, affodance:The avatar can speak to them by using the “chat_avatar” action with the object_id specified.}".

[0082] The "Selectable actions" section of this prompt contains information about the actions the autonomous agent can take. The action column for this prompt lists the following actions that the autonomous agent can take: "idle: Do nothing", "move_around: Walk around", "move_to_area: Move to another area", and "Interacting with objects: Actions included in the affordances of the object".

[0083] The output format "Output" for this prompt includes the output format for information about the autonomous agent's behavior generated by the large-scale language model 200. This output format for this prompt includes the reason "Reason" and the action "Action" as the output format for information about the autonomous agent's behavior generated by the large-scale language model 200.

[0084] The "Reason" in this prompt is the reason for generating the autonomous agent's action. The "Action" in this prompt is information about the generated autonomous agent's action.

[0085] Next, we will describe an example of the output of the large-scale language model 200. Figure 9 shows an example of the output of the large-scale language model. As shown in Figure 9, the output of the large-scale language model 200 includes reason "Reason" and action "Action".

[0086] The "Reason" for this output is the reason for the autonomous agent's action. In the "Reason" field for this output, the reason for the autonomous agent's action is stated as: "I am currently idle in shopping_area_01. As a salesperson, my role is to actively engage with players to sell products. There are two players (player_abatar_1 and player_abatar_2) present in the shopping area. Since I haven't interacted with any player yet, I should approach and engage one of them."

[0087] The "Action" in this output contains information about the actions of the generated autonomous agent. In the action column of this output, "chat_abatar with player_abatar_1" is listed as information about the actions of the generated autonomous agent.

[0088] Next, the effects and advantages of the embodiments of this disclosure will be described.

[0089] The autonomous agent action decision device 100 in the embodiment of this disclosure determines the autonomous agent's actions based on information about the area where the autonomous agent resides and information about objects present in that area. This allows the autonomous agent to perform a variety of actions. Furthermore, the autonomous agent action decision device 100 generates prompts for generating information about the autonomous agent's actions based on the information acquired by the acquisition unit. This allows the large-scale language model to be instructed to generate information about the autonomous agent's actions. Here, the information acquired by the acquisition unit, namely the information about the area and the information about objects present in that area, becomes the affordance of the information about the autonomous agent's actions generated by the large-scale language model. In this way, since prompts can be created based on affordances, the learning cost in prompt generation and the latency in determining the autonomous agent's actions can be reduced.

[0090] Furthermore, this device includes a reception unit that receives information identifying the autonomous agent and information regarding the autonomous agent's position within the virtual space. Therefore, the autonomous agent's actions can be easily determined.

[0091] Furthermore, this device stores information about multiple areas in an area database, and retrieves information about the areas where the autonomous agent is located based on the information about the autonomous agent's location and the information stored in the area database. Therefore, it is possible to easily obtain information about the areas where the agent is located.

[0092] Furthermore, this device stores information about objects present in each of multiple areas in its object database. Based on the autonomous agent's location and the information stored in the object database, it retrieves information about objects present in the area. Therefore, it can easily retrieve information about objects present in the area.

[0093] Furthermore, this device sends prompts to the generating AI to obtain information about the autonomous agent's behavior from the generating AI, and then determines the autonomous agent's actions based on the obtained information about the autonomous agent's behavior. Therefore, it can quickly determine the appropriate action for the autonomous agent.

[0094] Furthermore, this device determines the actions of autonomous agents based on information about their individual characteristics. Therefore, it can determine the actions of autonomous agents more appropriately.

[0095] Furthermore, this device stores information about the personality of autonomous agents in a personality database, and retrieves information about the personality of an autonomous agent based on the information that identifies the autonomous agent and the information stored in the personality database. Therefore, information about the personality of an autonomous agent can be easily obtained.

[0096] Furthermore, this device determines the actions of the autonomous agent based on information about the actions the autonomous agent can take. Therefore, it can determine the autonomous agent's actions more appropriately.

[0097] Furthermore, this device stores information about the actions that an autonomous agent can take in each of multiple areas in its behavior database, and retrieves information about the actions that an autonomous agent can take based on the information stored in the behavior database. Therefore, it is possible to easily obtain information about the actions that an autonomous agent can take.

[0098] The apparatus and method of this disclosure have the following configurations.

[0099] [1] An apparatus comprising: an acquisition unit that acquires information about an area in which an autonomous agent acting autonomously in the virtual space exists, and information about objects present in the area in which an autonomous agent acts autonomously in the virtual space, among a plurality of areas in the virtual space; a decision unit that determines the actions of the autonomous agent based on the information about the area in which an autonomous agent exists and the information about objects present in the area in which an autonomous agent exists; and a prompt generation unit that generates a prompt for generating information about the actions of the autonomous agent based on the information acquired by the acquisition unit.

[0100] [2] The apparatus according to [1], further comprising a receiving unit that receives information identifying the autonomous agent and information regarding the location of the autonomous agent in the virtual space.

[0101] [3] The apparatus according to [1] or [2], further comprising an area database storing information relating to each of the plurality of areas, wherein the acquisition unit acquires the information relating to the existing area based on information relating to the position of the autonomous agent and the information stored in the area database.

[0102] [4] The apparatus according to any one of [1] to [3], further comprising an object database storing information about objects present in each of the plurality of areas, wherein the acquisition unit acquires the information about the objects present in the presence area based on information about the position of the autonomous agent and the information stored in the object database.

[0103] [5] The apparatus according to any one of [1] to [4], further comprising a transmitting / receiving unit that transmits the prompt to a generating AI and obtains information from the generating AI regarding the actions of the autonomous agent, wherein the determination unit determines the actions of the autonomous agent based on the information regarding the actions of the autonomous agent obtained by the transmitting / receiving unit.

[0104] [6] The apparatus according to any one of [1] to [5], wherein the acquisition unit acquires information relating to the personality of the autonomous agent, and the determination unit determines the behavior of the autonomous agent based on the information relating to the personality of the autonomous agent.

[0105] [7] The apparatus according to [6], further comprising a personality database storing information relating to the personality of the autonomous agent, wherein the acquisition unit acquires information relating to the personality of the autonomous agent based on information identifying the autonomous agent and information stored in the personality database.

[0106] [8] The apparatus according to any one of [1] to [7], wherein the acquisition unit acquires information relating to actions that the autonomous agent can take, and the decision unit determines the actions of the autonomous agent based on the information relating to actions that the autonomous agent can take.

[0107] [9] The apparatus according to [8], further comprising an action database storing information on actions that the autonomous agent may take in each of the plurality of areas, wherein the acquisition unit acquires the information on actions that the autonomous agent may take based on the information stored in the action database.

[0108]

[10] A method comprising: an acquisition step of acquiring information about an area in which an autonomous agent acting autonomously in the virtual space exists, and information about objects present in the area in which an autonomous agent acts autonomously in the virtual space, among a plurality of areas in the virtual space; a decision step of determining the actions of the autonomous agent based on the information about the area in which an autonomous agent exists and the information about objects present in the area in which an autonomous agent exists; and a prompt generation step of generating a prompt for generating information about the actions of the autonomous agent based on the information acquired in the acquisition step.

[0109] The block diagram used in the description of the above embodiment shows functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or it may be realized using two or more physically or logically separated devices that are directly or indirectly connected (for example, using wired or wireless connections). A functional block may be realized by combining the one or more devices with software.

[0110] Functions include, but are not limited to, judgment, decision, determination, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, assumption, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), and assigning. For example, a functional block (configuration part) that enables transmission is called a transmitting unit or transmitter. In all cases, as mentioned above, the method of implementation is not particularly limited.

[0111] For example, the autonomous agent action decision device 100 in one embodiment of the present disclosure may function as a computer that processes the autonomous agent action decision of the present disclosure. Figure 10 is a diagram showing an example of the hardware configuration of the autonomous agent action decision device 100 according to one embodiment of the present disclosure. The above-described autonomous agent action decision device 100 may be physically configured as a computer device including a processor 1001, memory 1002, storage 1003, communication device 1004, input device 1005, output device 1006, bus 1007, etc.

[0112] In the following explanation, the term "device" can be replaced with "circuit," "device," "unit," etc. The hardware configuration of the autonomous agent action decision device 100 may include one or more of the devices shown in the figure, or it may be configured to omit some of the devices.

[0113] Each function in the autonomous agent action decision device 100 is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, which allows the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of data reading and writing in the memory 1002 and storage 1003.

[0114] The processor 1001 controls the entire computer, for example, by running the operating system. The processor 1001 may be composed of a central processing unit (CPU) that includes interfaces with peripheral devices, control devices, arithmetic units, registers, etc. For example, the decision unit 103 and the prompt generation unit 104 described above may be implemented by the processor 1001.

[0115] Furthermore, the processor 1001 reads programs (program code), software modules, data, etc., from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes accordingly. The program used is one that causes the computer to execute at least a part of the operations described in the above embodiment. For example, the determination unit 103 and the prompt generation unit 104 may be implemented by a control program stored in the memory 1002 and running on the processor 1001, and other functional blocks may be implemented similarly. The above-described various processes have been explained as being executed by one processor 1001, but they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may also be transmitted from a network via a telecommunications line.

[0116] The memory 1002 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. The memory 1002 may also be called a register, cache, main memory, etc. The memory 1002 can store executable programs (program code), software modules, etc., for carrying out autonomous agent action decisions according to one embodiment of the present disclosure.

[0117] The storage 1003 is a computer-readable recording medium and may consist of at least one of the following: an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disc, a digital multipurpose disc, a Blu-ray® disc), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. The storage 1003 may also be called an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, server, or other suitable medium including at least one of memory 1002 and storage 1003.

[0118] The communication device 1004 is hardware (transmitting / receiving device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc. The communication device 1004 may be configured to include high-frequency switches, duplexers, filters, frequency synthesizers, etc., in order to implement at least one of frequency division duplex (FDD) and time division duplex (TDD). For example, the receiving unit 101, transmitting / receiving unit 105, and request unit 106 described above may be implemented by the communication device 1004. The communication device 1004 may be implemented with physically or logically separated transmitting and receiving units.

[0119] The input device 1005 is an input device that accepts input from an external source (e.g., a keyboard, mouse, microphone, switch, button, sensor, etc.). The output device 1006 is an output device that outputs to an external source (e.g., a display, speaker, LED lamp, etc.). The input device 1005 and the output device 1006 may be configured as an integrated unit (e.g., a touch panel).

[0120] Furthermore, each device, such as the processor 1001 and the memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or different buses may be configured for each device.

[0121] Furthermore, the autonomous agent action decision device 100 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array), and some or all of each functional block may be realized by such hardware. For example, the processor 1001 may be implemented using at least one of these hardware components.

[0122] Information notification is not limited to the embodiments described herein and may be carried out by other means. For example, information notification may be carried out by physical layer signaling (e.g., DCI (Downlink Control Information), UCI (Uplink Control Information)), upper layer signaling (e.g., RRC (Radio Resource Control) signaling, MAC (Medium Access Control) signaling, broadcast information (MIB (Master Information Block), SIB (System Information Block))), other signals, or combinations thereof. RRC signaling may also be called RRC messages, and may be, for example, RRC Connection Setup messages, RRC Connection Reconfiguration messages, etc.

[0123] The processing procedures, sequences, flowcharts, etc., of each aspect / embodiment described in this disclosure may be reordered, provided they do not contradict each other. For example, the methods described in this disclosure present various step elements using exemplary order and are not limited to the specific order presented.

[0124] Input and output information may be stored in a specific location (e.g., memory) or managed using a management table. Input and output information may be overwritten, updated, or appended to. Output information may be deleted. Input information may be transmitted to other devices.

[0125] The determination may be made by a value represented by one bit (0 or 1), by a boolean value (true or false), or by a numerical comparison (for example, a comparison with a predetermined value).

[0126] Each aspect / embodiment described in this disclosure may be used individually, in combination, or switched between as needed during implementation. Furthermore, notification of specific information (e.g., notification that "X is") is not limited to explicit notification, but may also be implicit (e.g., by not providing such notification).

[0127] Although the present disclosure has been described in detail above, it will be clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the intent and scope of the present disclosure as defined by the claims. Accordingly, the descriptions in the present disclosure are illustrative and not intended to be restrictive in any way.

[0128] Software should be broadly interpreted to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, execution threads, procedures, functions, and so on, whether they are called software, firmware, middleware, microcode, hardware description languages, or by any other name.

[0129] Furthermore, software, instructions, information, etc., may be transmitted and received via a transmission medium. For example, if software is transmitted from a website, server, or other remote source using at least one of wired technologies (such as coaxial cable, fiber optic cable, twisted pair, or digital subscriber line (DSL)) and wireless technologies (such as infrared or microwave), then at least one of these wired and wireless technologies is included in the definition of a transmission medium.

[0130] The information, signals, etc. described in this disclosure may be represented using any of the various different techniques. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

[0131] In addition, terms used in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings. For example, at least one of the channel and symbol may be a signal (signaling). Also, a signal may be a message. Furthermore, a component carrier (CC) may be called a carrier frequency, cell, frequency carrier, etc.

[0132] Furthermore, the information, parameters, etc., described in this disclosure may be expressed using absolute values, relative values ​​from a given value, or other corresponding information. For example, wireless resources may be indicated by an index.

[0133] The names used for the parameters described above are not restrictive in any way. Furthermore, the formulas and other expressions using these parameters may differ from those expressly disclosed in this disclosure. Various channels (e.g., PUCCH, PDCCH, etc.) and information elements can be identified by any suitable name, and therefore, the various names assigned to these various channels and information elements are not restrictive in any way.

[0134] In this disclosure, terms such as "Mobile Station (MS)," "user terminal," "User Equipment (UE)," and "terminal" may be used interchangeably.

[0135] A mobile station may also be referred to by those skilled in the art as a subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal, handset, user agent, mobile client, client, or some other appropriate term.

[0136] As used in this disclosure, the terms “determining” and “determining” may encompass a wide variety of actions. “Determining” may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, or inquiring (e.g., searching in a table, database, or other data structure), or ascertaining. “Determining” may also include, for example, receiving (e.g., receiving information), transmitting (e.g., sending information), inputting, outputting, or accessing (e.g., accessing data in memory). Furthermore, "judgment" and "decision" can include considering something as having been "judged" or "decided" after resolving, selecting, choosing, establishing, comparing, etc. In other words, "judgment" and "decision" can include considering something as having been "judged" or "decided" after some action. Also, "judgment (decision)" can be reinterpreted as "assuming," "expecting," or "considering."

[0137] The terms “connected,” “coupled,” or any variation thereof, mean any direct or indirect connection or coupling between two or more elements, and may include the presence of one or more intermediate elements between two elements that are “connected” or “coupled” with each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, “connection” may be reinterpreted as “access.” As used in this disclosure, two elements may be considered to be “connected” or “coupled” with each other using at least one of one or more wires, cables, and printed electrical connections, and, in some non-limiting and non-exclusive examples, electromagnetic energy having wavelengths in the radio frequency domain, microwave domain, and optical (both visible and invisible) domain.

[0138] In this disclosure, the phrase "based on" does not mean "based solely on" unless otherwise specified. In other words, the phrase "based on" means both "based solely on" and "based at least on."

[0139] Any reference to elements using designations such as “first,” “second,” etc., as used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Accordingly, references to first and second elements do not imply that only two elements may be employed, or that the first element must precede the second element in any way.

[0140] Where the terms “include,” “including,” and their variations are used in this disclosure, these terms are intended to be inclusive, as is the term “comprising.” Furthermore, the term “or” as used in this disclosure is not intended to be exclusive OR.

[0141] In this disclosure, if articles are added by translation, such as a, an, and the in English, this disclosure may include the fact that the noun following these articles is plural.

[0142] In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combine" may be interpreted similarly to "different."

[0143] 100...Autonomous agent action decision device, 101...Reception unit, 102...Acquisition unit, 103...Decision unit, 104...Prompt generation unit, 105...Transmission / reception unit, 106...Request unit, 200...Large-scale language model, 300...Game server, 400...Area database, 500...Object database, 600...Personality database, 700...Action database, 800...Data parser, 900...User terminal, 1001...Processor, 1002...Memory, 1003...Storage, 1004...Communication device, 1005...Input device, 1006...Output device, 1007...Bus.

Claims

1. An apparatus comprising: an acquisition unit that acquires information about an area in which an autonomous agent acting autonomously in the virtual space exists, and information about objects present in the area in which the autonomous agent exists, from among a plurality of areas in the virtual space; a decision unit that determines the actions of the autonomous agent based on the information about the area in which the autonomous agent exists and the information about objects present in the area in which the autonomous agent exists; and a prompt generation unit that generates a prompt for generating information about the actions of the autonomous agent based on the information acquired by the acquisition unit.

2. The apparatus according to claim 1, further comprising a receiving unit that receives information identifying the autonomous agent and information regarding the position of the autonomous agent in the virtual space.

3. The apparatus according to claim 1, further comprising an area database storing information relating to each of the plurality of areas, wherein the acquisition unit acquires the information relating to the area based on information relating to the location of the autonomous agent and the information stored in the area database.

4. The apparatus according to claim 1, further comprising an object database storing information about objects present in each of the plurality of areas, wherein the acquisition unit acquires the information about the objects present in the presence area based on information about the position of the autonomous agent and the information stored in the object database.

5. The apparatus according to claim 1, further comprising a transmitting / receiving unit that transmits the prompt to a generating AI and obtains information from the generating AI regarding the behavior of the autonomous agent, wherein the determination unit determines the behavior of the autonomous agent based on the information regarding the behavior of the autonomous agent obtained by the transmitting / receiving unit.

6. The apparatus according to claim 1, wherein the acquisition unit acquires information relating to the personality of the autonomous agent, and the determination unit determines the behavior of the autonomous agent based on the information relating to the personality of the autonomous agent.

7. The apparatus according to claim 6, further comprising a personality database storing information relating to the personality of the autonomous agent, wherein the acquisition unit acquires information relating to the personality of the autonomous agent based on information identifying the autonomous agent and information stored in the personality database.

8. The apparatus according to claim 1, wherein the acquisition unit acquires information relating to actions that the autonomous agent can take, and the decision unit determines the action of the autonomous agent based on the information relating to actions that the autonomous agent can take.

9. The apparatus according to claim 8, further comprising an action database storing information relating to actions that the autonomous agent may take in each of the plurality of areas, wherein the acquisition unit acquires the information relating to actions that the autonomous agent may take based on the information stored in the action database.

10. A method comprising: an acquisition step of acquiring information about an area in which an autonomous agent acting autonomously in the virtual space exists, and information about objects present in the area in which the autonomous agent exists, from among a plurality of areas in the virtual space; a decision step of determining the actions of the autonomous agent based on the information about the area in which the autonomous agent exists and the information about objects present in the area in which the autonomous agent exists; and a prompt generation step of generating a prompt for generating information about the actions of the autonomous agent based on the information acquired in the acquisition step.

Images