Information processing system, program, and information processing method

Abstract

To provide an information processing system, a program, and an information processing method capable of enhancing a player's interest in a game using a non-player's character (NPC).SOLUTION: An information processing system for providing a game using a non-player's character includes: an NPC control section for controlling a non-player's character on the basis of at least one learned model; an additional information acquisition section for acquiring setting information associated with a game medium; and a model generation section for generating a learned model with a setting on the basis of setting information associated with a game medium acquired by the additional information acquisition section. The NPC control section controls the non-player's character by reflecting the setting information on the basis of the learned model with a setting.SELECTED DRAWING: Figure 5A

Description

[Technical field]

[0001] The present disclosure relates to an information processing system, a program, and an information processing method. [Background technology]

[0002] As a game that can be executed on player devices such as smartphones and game consoles, There are known games that use non-player characters (NPCs) controlled by a computer. is.

[0003] For example, Patent Document 1 discloses a system that allows players to converse with NPCs that appear in a game. In the game of Patent Document 1, the player and the NPC exchange The contents of the conversation are stored in the game system as text data beforehand, and the player The conversation with the NPC progresses when Iya selects a line from a selection of lines. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP2010-246758A Summary of the Invention [Problem to be solved by the invention]

[0005] In the game system of Patent Document 1, the conversation progress and the emotion level of the NPC are associated with each other. There are multiple lines prepared, and the next conversation will be decided according to the player's choice of lines. It is possible to change the content and emotions of NPCs in conversation. The changes in the content of the dialogue are limited to the range of pre-prepared dialogue options. If we can diversify the expressions related to the game, we can expect the game to become more interesting. can.

[0006] The objective of the exemplary embodiment of the present disclosure is to develop a game using non-player characters (NPCs). Information processing system and program capable of improving the interest of players in a game The present invention provides a system and an information processing method. [Means for solving the problem]

[0007] An information processing system according to one aspect of the present disclosure includes: An information processing system for providing a game using a non-player character, comprising: N that controls the non-player characters based on at least one learned model. A PC control unit, an additional information acquisition unit that acquires setting information associated with a game medium; Based on the setting information associated with the game medium acquired by the additional information acquisition unit, A model generation unit that generates a trained model, The NPC control unit performs the above-mentioned processing based on the learned model with the setting, while reflecting the setting information. Controlling a non-player character.

[0008] By having the above-mentioned features, the information processing system can play games using non-player characters. This can increase the player's interest in the game.

[0009] A program according to one aspect of the present disclosure includes: Executing information processing related to the provision of games using non-player characters on a computer A program for causing Controlling the non-player character based on at least one learned model. And, Obtaining setting information associated with a game medium; A trained model with settings is generated based on the setting information associated with the acquired game content. To do, The setting information is used to control the non-player character based on the learned model with the setting. and causing the computer to execute the steps of:

[0010] By having the above features, the program can be used in games that use non-player characters. This can increase the interest of players.

[0011] An information processing method according to one aspect of the present disclosure includes: A computer executes information processing related to the provision of a game using non-player characters. An information processing method comprising: Controlling the non-player character based on at least one learned model. And, Obtaining setting information associated with a game medium; A trained model with settings is generated based on the setting information associated with the acquired game content. To do, The setting information is used to control the non-player character based on the learned model with the setting. and reflecting the result in the computer.

[0012] By having the above-mentioned characteristics of the information processing method, in a game using a non-player character, This can increase the interest of the player. [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 is a schematic diagram illustrating a configuration example of an information processing system according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a block diagram illustrating an example of a functional configuration of the server illustrated in FIG. [Diagram 3] FIG. 3 is a block diagram showing an example of a functional configuration of the player terminal shown in FIG. [Figure 4] FIG. 4 is a schematic diagram illustrating an example of game media information stored in the game information storage unit illustrated in FIG. [Figure 5A] FIG. 5A is a flowchart for explaining an outline of the processing of the information processing system. [Figure 5B] FIG. 5B is a flowchart for explaining an outline of the processing of the information processing system. [Figure 5C] FIG. 5C is a flowchart for explaining an outline of the processing of the information processing system. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] An information processing system 10 according to an embodiment of the present disclosure will be described with reference to the drawings. The drawings are merely examples for explaining the present embodiment. The diagrams and flow charts may be modified or changed within the scope that does not cause technical problems. In addition, in the present specification and drawings, The same reference numerals are used to denote the same components, and duplicate explanations will be omitted.

[0015] The information processing system 10 according to the present embodiment is connected to a network N such as the Internet. This is a system for providing various game-related services to player 3 via the above.

[0016] As shown in FIG. 1, the information processing system 10 includes at least one server 20 and a few The main components of the game include at least one player terminal 30. In this information processing system 10, A server 20 provided by a company 1 provides a game that includes images, videos, sounds, etc. By having the player terminal 30 owned by the player 3 execute information processing related to the A game is provided to the player terminal 30.

[0017] <Server device configuration> The server 20 is an information processing device used to operate and manage the game service. As described above, the information processing device is provided to a company 1 that provides game services. The server 20 may be, for example, a workstation or a personal computer. It can be a general-purpose computer, and the cloud computing technology can be used to The structure may be constructed in a systematic way.

[0018] The server 20 includes a control unit 21 that controls the entire server 20, and a controller 22 that controls various data and applications. A storage unit 22 for storing an action program, and a player terminal 30 connected via a network N. and a server communication unit 23 for transmitting and receiving data between the server and the server.

[0019] <Control unit (server)> The control unit 21 transfers data between each unit and controls the entire server 20. The control unit 21 executes various information processes required for providing game services. The function of the control unit 21 is implemented by a CPU (Central Processor). ing Unit) and / or GPU (Graphics Processing Unit) A processing unit such as a computing device processes data and programs stored in a specified storage device. This is achieved by reading it out and executing various programs in the working area of ​​the memory.

[0020] The control unit 21 has a number of functional units corresponding to the types of information processing to be executed. The game execution unit 210, the NPC control unit 211, and the additional information acquisition unit 212 are functional units. a processing result generating unit 215, a processing result generating unit 216, a processing result receiving unit 217, a processing result generating unit 218, a processing result generating unit 219, a processing result generating unit 220, a processing result generating unit 221, a processing result generating unit 222, a processing result generating unit 223, a processing result generating unit 224, a processing result generating unit 225, a processing result generating unit 226, a processing result generating unit 2 , may be provided.

[0021] The game execution unit 210 executes a process for progressing the game according to the game program. The categories of games provided by the information processing system 10 are not particularly limited. For example, love games, battle games, puzzle games, action games, baseball games, soccer games, etc. Car games, other sports games, quiz games, pinball games, card games, Rhythm games, RPGs, adventure games, casino games, simulation games , strategy games, board games, racing games, etc.

[0022] Games also include, for example, so-called "in-game" content related to actual gameplay. It is a part where you control characters in the game or puzzles to solve them. You can progress through stages (quests), play cards, race against other players, etc. In addition to the "game progression part" to progress the game, ", in-game characters, items, and other game media are given to players through a lottery. The "lottery part" such as the so-called "gacha" that allows players to select characters and other game media used in the game. The game includes a "formation part" where you select your character, and a "strengthening part" where you can strengthen your characters and items. It's okay to do that.

[0023] The game in which the game execution unit 210 progresses includes non-player characters (hereinafter referred to as NPCs). NPCs are characters that are not controlled by the player. For example, NPCs are controlled by an artificial intelligence called character AI. The use of the is not particularly limited. For example, in a battle game, the player and the teammate who cooperate with each other can be You can use NPCs as characters, or NPs as enemy characters to fight against the player. C may be used. In addition, in the case of a romance game, an NPC may be used as the love interest. In the case of a training-type simulation game, N characters may be used as the characters to be trained. A PC may be used. In a game in which the game execution unit 210 advances, a single NPC is used. Alternatively, multiple NPCs may be used.

[0024] In addition, the game in which the game execution unit 210 progresses uses at least one game medium. Here, the "game media" refers to electronic data that the player 3 uses in the game. For example, cards, characters that are the objects of operation of the player 3 (hereinafter, the operation (sometimes referred to as "characters"), avatars, in-game currency and other consumable media, and Other items and the like correspond to game media. The player 3 operates an avatar or a character that operates in the game space to play the game. In addition, weapons, armor, clothing, etc. may be used as game media in the game. Items such as clothing and accessories are added to the avatar of player 3 according to the selection of player 3. It can also be equipped to the player character or NPC. As game media, interior items and other items are added to the game space according to the selection of the player 3. A game may be executed by placing the device between the two devices.

[0025] The functions of the NPC control unit 211, the additional information acquisition unit 212, and the model generation unit 213 are This corresponds to the function of character AI.

[0026] The NPC control unit 211 performs at least Each has the function of executing the process to control the NPC based on a single trained model. That is, the NPC control unit 211 is the main component of the character AI. The trained model used in the processing is a machine learning model that allows NPCs to act autonomously. This is a decision-making model for NPCs. In the NPC control unit 211, for example, the player 3 decides to prompts sent to player 3, other operation information sent by player 3 before the game starts, The information to be set (for example, the information to be used in the game determined by the selection operation of the player 3) Information on the game media to be played, information on the game units to be played (quests, events, etc.), etc. ), information about the game situation collected during the game, and past At least one piece of data selected from the game history, etc. is used as input to the trained model. The NPC control unit 211 inputs the above data into the trained model and Determines and outputs the actions (words and / or actions) that the NPC will perform in the game. The control unit 211 receives control information for the NPC, which is the output of the trained model (words to be executed by the NPC ( The processing result generating unit 215 transmits the information (such as conversation and / or action information) to the processing result generating unit 215, which will be described later.

[0027] The information processing system 10 is a single system that comprehensively controls the activities of NPCs in the game space. A trained model may be used, or multiple trained models corresponding to specific control items may be combined. They may be used together (i.e., multiple trained models may be integrated to control NPCs) For example, the NPC control unit 211 controls both the words (conversation) and the actions of the NPC. The NPC may be controlled based on a trained model that outputs The NPC is generated based on a combination of the trained model and the trained model that outputs the NPC's behavior. In addition, the information processing system 10 may also use a trained model for controlling the NPC. As a rule, a common trained model may be prepared for use by all players 3, or For each layer 3, a trained model incorporating (trained) information associated with each individual player 3 is created. A rule may be prepared.

[0028] Machine learning to build a trained model is necessary to realize the autonomous actions of NPCs. The machine learning method can be implemented by selecting the appropriate learning data and learning algorithm. There is no particular limitation. Machine learning may be unsupervised learning or supervised learning. It may be a learning method, a reinforcement learning method, or a deep learning method. These machine learning techniques may be combined to build a trained model.

[0029] The training data is not particularly limited and may be appropriately selected so as to obtain a desired output. For example, the training data for outputting NPC words (generating conversations) using a model is as follows: A large-scale corpus of natural language sentences, natural language input (prompts) and corresponding inputs A conversation data set consisting of a combination of human responses to the input and an evaluation dataset in which humans evaluated the appropriateness of responses to prompts. In addition, the trained model outputs both the NPC's conversation and the actions taken when talking. In this case, in addition to the training data, the natural language and gestures used to utter that natural language are also required. A dataset consisting of combinations of character patterns and human evaluations of the appropriateness of the behavior. In addition, when outputting the actions of NPCs in a battle game, etc. In this case, for example, data from a fighting game in which a person operates a character can be used as learning data. In addition to the learning data described above, game specifications and in-game rules may also be used. The model may be trained to learn basic information about the game.

[0030] The learning algorithm may be appropriately selected so as to obtain the desired output. There is no limitation, and any known learning algorithm can be used. For example, deep neural network neural networks, convolutional neural networks, recurrent neural networks, etc. Which neural network, RLFH (Reinforcement Learning g from Human Feedback), SFT (Supervised Fi Reinforcement learning algorithms such as Ne-Tuning, decision trees, random forests, and One or more algorithms selected from the group consisting of neural network analysis, principal component analysis, etc. may be used. It is preferable to employ a network.

[0031] In the information processing system 10 according to the present embodiment, information is sent to the character AI that controls the NPC. In this embodiment, the following is executed to add information to the trained model for controlling the NPC: The existing model before the information was added is called the "predecessor model," and the model after the information was added is called the "predecessor model." The latter model will be called the "pre-trained model with settings." The information added to this document is referred to as "Additional Information."

[0032] The NPC control unit 211 receives a trained model with settings generated by the model generation unit 213 described later. When the NPC is notified of the additional information, the NPC is controlled based on the learned model with the settings. In other words, when a learned model with settings is generated, the NPC control unit 211 Changed the control of the PC from the control based on the predecessor model to the control based on the trained model with settings. When executing the output reflecting the additional information, the NPC control unit 211 The trained model used to control the NPC (hereinafter, sometimes referred to as the control model) is Alternatively, you can execute a process to switch from the self model to a trained model with settings. The C control unit 211 controls the NPC by using the predecessor model and the learned model with settings in combination. Good too.

[0033] In the following explanation, we will assume that the game contains scenes in which NPCs converse, and assume that character A This section will illustrate the process of adding information to I. In other words, a trained model with settings, etc. This mainly illustrates the case where a trained model generates NPC dialogue based on additional information. However, a trained model such as a trained model with settings may not be able to use control information related to the conversation. The model is not limited to outputting control information related to the activities of NPCs in the game. For example, a pre-trained model with settings can be used to train the NPC's combat activities. The model may output control information that allows the NPC to talk to the player. It may be a conversation between three NPCs, or it may be a conversation between one NPC and another NPC. In addition, conversations between Player 3 and NPCs are conducted by Player 3 through an avatar or a controlled character. It may be a conversation with an NPC, or the player 3 may use a medium such as an avatar or a controlled character to The conversation may be conducted without the intervention of the NPC. In the case of a model that outputs control information related to the player, the pre-trained model with settings includes the conversation of player 3. is input (which may be in text form or voice form), and Information generated during the game may also be input.

[0034] The additional information acquisition unit 212 has a function of executing a process for acquiring setting information associated with the game medium. The setting information associated with the game media is incorporated into the character AI that controls the NPC. This corresponds to additional information. In other words, in this embodiment, data to be additionally trained on the trained model (Additional learning data) is managed by linking it to the game media that the player 3 can use in the game. In addition to the setting information associated with the game media, the trained model also contains a list of game media, Other game-related information, such as the past game history of Player 3, may be additionally learned. stomach.

[0035] The game medium used in the information processing system 10 includes setting information that serves as additional learning data. It includes linked game media, and also links to setting information that serves as additional learning data. In the following description, the setting that is the additional learning data may include game content that is not included in the game content. When indicating whether information is present or not, configuration information that serves as additional learning data is linked. The game medium is referred to as the "first game medium," and other game media that are not linked to the setting information are referred to as the "second game medium." This will be called the "second game medium." A pre-trained model with settings will be used to generate NPC conversations. If you are using a character, you will need to add an item to the character (avatar, player character, NPC, etc.). The game includes items placed in the game space, items consumed during game progress, and NPCs. Any character or the like may be set in the first game medium.

[0036] In the game provided by the information processing system 10, the method of acquiring the game contents is For example, the game medium may be a lottery method such as gacha, or a reward for clearing a game. rewards, exchange for paid or free consumable media to be consumed in the game, and exchange for two or more objects The player 3 may be given the power by a method such as synthesis of the object. An NPC that is not a member of the first game medium may be given to the player 3 in the above manner. The method of winning may be subject to special restrictions different from those of the second game medium. For example, The medium is given to the player 3 by a method such as a limited-time gacha or a reward for a limited-time event. In addition, the information may be registered from the beginning of the operation of the game system (the beginning of the game provision). The game medium that is added to the game system later is called the second game medium, and the game medium that is added to the game system later is called the first game medium. As a game medium, setting information may be linked to the game medium.

[0037] The additional information acquisition unit 212 is triggered by the acquisition of the first game medium by the player 3. The processing may include acquiring setting information associated with the first game medium. The acquisition unit 212 detects that the player 3 has acquired the first game medium and triggers the detection. The game controller 100 then acquires setting information associated with the first game medium. are transmitted to the model generation unit 213, which will be described later, as data for additional learning.

[0038] The trigger for starting the processing of the additional information acquisition unit 212 is the above-mentioned For example, the additional information acquisition unit 212 may acquire the first game medium by the player 3. A process for acquiring setting information associated with the first game medium is executed using the use of the body as a trigger. The player 3 may select a certain number of players during the preparation process before the game is executed and / or during the game progress. Select the game media to be used in the game (referred to as the "used game media") from among the available game media. In this case, the additional information acquisition unit 212 acquires the additional information of the game media selected by the player 3. Identifying a first game medium from the list, and using the identification as a trigger, setting associated with the first game medium. The acquired setting information may be used as additional learning data to be described later. The model generation unit 213 then transmits the model to the model generation unit 213.

[0039] The additional information acquisition unit 212 also acquires updates to the game system provided by the server 20. A process to obtain additional information may be executed using the update as a trigger. When a new game medium is registered in the storage unit 22 due to a transfer, the additional information acquisition unit 212 Identifying a first game medium from among the newly registered game media and linking the first game medium The acquired setting information may be used as additional learning data. , and transmits it to the model generation unit 213 described later.

[0040] The model generation unit 213 generates the setting associated with the first game medium acquired by the additional information acquisition unit 212. It has the function of executing the process of generating a trained model with settings based on information (additional information). In other words, the model generation unit 213 converts the setting information associated with the first game medium into the learned model. A process of additionally learning the model is performed. A plurality of trained models are registered in the information processing system 10. If so, the model generation unit 213 performs the following in accordance with the setting information associated with the first game medium: A process may be performed to select a trained model to which the configuration information is added.

[0041] In the additional learning performed by the model generation unit 213, only the setting information associated with the first game medium is The setting information associated with the first game medium may be trained in a trained model, or the setting information associated with the first game medium may be trained in trained data. The trained model may be retrained in combination with the above. The method of additional training is not particularly limited. For example, reinforcement learning, fine-tuning, or transfer learning may be employed. When adding setting information associated with the first game medium in tuning, for example, A part or the whole of the model is reproduced using learning data including setting information associated with the first game medium. The weights of the entire network of the model are updated by learning. When adding configuration information associated with the output layer of the trained model network, The output layer is then trained on additional information, including setting information associated with the first game medium. The weights of the learning algorithm used in fine tuning or transfer learning are updated. A publicly known model may be used. In addition, when generating a trained model with settings, additional training may be performed. The predecessor model before the configuration can be left as it is, or the predecessor model can be updated to perform the configuration-based learning. A completed model may be generated.

[0042] When the learned model with settings is generated by the model generation unit 213, the NPC control unit 2 11 will shift the control of NPCs from a predecessor model to a pre-trained model with settings. This change in control model is automatically made when a trained model with settings is generated. Alternatively, the NPC control unit 211 may execute the first game medium by the player 3. The acquisition or use of the body is detected, and the detection is used as a trigger to A change to the control to be used may be executed. That is, the NPC control unit 211 and detecting acquisition or use of a first game medium by the user, and acquiring setting information associated with the first game medium. For example, the additional information acquisition unit 212 may reflect the activity of the NPC in the game. The update of the game system is used as a trigger to acquire setting information associated with the first game medium. In this case, the NPC control unit 211 may notify the player 3 of the acquisition or use of the first game medium. The control model may be changed as a trigger. Setting associated with the first game medium, triggered by acquisition of the first game medium by layer 3 When the information is acquired, the NPC control unit 211 notifies the player 3 of the use of the first game medium. The control model may be changed using the use of the controller as a trigger.

[0043] The change in the control model may be performed based on a selection made by the user. For example, By additionally learning the setting information linked to the medium, it is possible to generate conversations in dialects. When a model is generated, the NPC's conversation is controlled in standard mode based on the predecessor model. It may be controlled in dialect mode based on a pre-trained model with settings, and the user may select a standard By selecting either the language mode or the dialect mode, you can switch the control of the NPC. It is also possible to change the control of NPCs from a predecessor model to a trained model with settings. After changing the control to the one shown in FIG. 1, the NPC control unit 211 monitors the progress of the game, the game used by the user, and the like. The control of NPCs is based on the selection of game media or user selection, etc., in the predecessor model. In this way, the control by the predecessor model and the control by the setting model can be restored. Even if it is possible to switch between control using a trained model and control using a non-trained model at a specified timing, good.

[0044] The operation information receiving unit 214 has a function of receiving operation information transmitted from the player terminal 30. The received operation information is used to process the game program by the game execution unit 210. The received operation information is used as instruction information for the NPC control unit 211. The input to the control model (predecessor model, trained model with settings) when controlling the activity of the NPC Used as prompt.

[0045] The processing result generating unit 215 generates the processing result of the game program executed by the game executing unit 210. The game program has a function of executing a process to generate a processing result. In this case, the processing result generation unit 215 receives the control information of the NPC from the NPC control unit 211. The information is received, and the processing result of the game program is generated by reflecting the control information. The processing results generated by the result generating unit 215 are transmitted to the player terminal via the server communication unit 23. The video signal is transmitted to the player terminal 30 and output by the display unit 34 of the player terminal 30.

[0046] If the trained model with settings outputs control information related to the NPC's conversation, The output on the display unit 34 derived from the information may be in text format. The control information related to the conversation may be output in the form of audio, such as the speech of an NPC. In addition to language information, NPC action information accompanying the conversation may also be included. If the output (control information) of the model includes information on the NPC's actions accompanying the conversation, the NPC's actions are However, the motion information may be output to the display unit 34 (for example, in accordance with the dialogue of the NPC). The mouth movement of the NPC saying "Yeeshaa!" is output. (The command will output "yes", etc.)

[0047] <Storage unit (server)> The memory unit 22 stores programs for executing various control processes and functions in the control unit 21, This is the storage area of ​​the server 20 that stores the force data, etc., and is RAM (Random Access Memory). s Memory), ROM (Read Only Memory), SRAM (Sta tic Random Access Memory), flash memory, HDD (H ard Disk Drive), SSD (Solid State Drive), etc. or any combination of other storage devices. The storage unit 22 temporarily stores the contents of communication by the server 20 when the server 20 performs communication. The present invention may have a function of:

[0048] The storage unit 22 of the server 20 has a plurality of areas partitioned according to the type of information to be stored. For example, the storage unit 22 may include a game information storage unit 221, a player information storage unit 2 22 and a model storage unit 223.

[0049] The game information storage unit 221 stores basic information about the game. The basic information about the game stored in the game information storage unit 221 is For example, the game information storage unit 221 stores a game unit Setting information for each rank, information about the game medium, information related to the game space such as background images, events In addition, a "game unit" refers to one game. A "game unit" refers to a collection of quests, events, missions, courses, and so on. Terms such as story, scenario, chapter, dungeon, stage, battle, and match This can be rephrased as:

[0050] The information on the game content stored in the game information storage unit 221 includes information that uniquely identifies each game content. An ID (hereinafter referred to as a game media ID) may be assigned as an identifier for identifying the game. In the game information storage unit 221, information on the game medium is stored, such as the name of the game medium, image data, etc. The data, acquisition method, etc. may also be stored.

[0051] FIG. 4 illustrates an example of information on the first game medium stored in the game information storage unit 221. FIG. 4 shows seven examples of first game media, of which M-A001 to M-A00 The type of the first game medium indicated by ID 6 is an item, and the type indicated by ID M-C001 is The type of first game media shown is a character.

[0052] As shown in FIG. 4, the information on the first game medium includes setting information linked to a game medium ID. (Additional learning data). The setting information associated with the first game medium is All that is required is additional learning data used to build a trained model, and the details of the setting information are There is no particular limitation. The trained model with settings is considered as a model that outputs control information related to the conversation. In this case, the setting information may be data such as those shown in (1) to (7) below (see FIG. 4). (See reference). Any combination of the data shown in (1) to (7) is set in the first game medium. The information may be linked.

[0053] (1) An episode (scenario) related to the item (2) Data concerning the character's experiences or career (3) Data regarding the character's status (status, position, occupation, position, etc.) (4) Character personality data (Big Five Personality onality traits) (5) Data regarding character preferences (6) The way the character speaks (intonation, accent, dialect, endings, etc.) Data involved (7) Data representing the relationship between characters

[0054] The data exemplified in (2) to (7) above is data associated with the first game medium, which is a character. The data may be game data, or may be data linked to the first game medium which is an item. If (2) to (7) are data to be linked to an item, then (2) to (7) are the “character "Item" refers to a character that is attached with an item, and the item is attached to an avatar or an operating character. It may be attached to an object controlled by the player 3, such as a robot, or to an NPC.

[0055] The data shown in (1) to (7) above are examples of setting information mainly related to conversation generation. However, when generating actions such as NPC combat actions, game media such as items are used. You can also assign settings to the body that relate to the NPC's actions. For example, you can assign a weapon to a specific item. Link the setting information about the attack skill of the armor item, and set the selection probability of the defense action and the attack One possibility would be to link it to the probability of avoidance, etc.

[0056] The information on the first game medium stored in the game information storage unit 221 includes information on the target game medium to be additionally learned. Information for identifying the elephant's trained model may be included. For example, " The "Target Model" column contains information to identify the trained model to be trained. The "conversation" shown in the target model column indicates that the trained model for conversation generation is the target for additional training. "Conversation and Action" means that the trained model for conversation generation and the trained model for action generation are set. Add one or more of the following: a trained model, a trained model that generates both conversations and actions, This means that the model is specified as a learning target. When selecting a trained model to add setting information from among the above, The trained model to be additionally trained may be determined based on information specifying the target model to be additionally trained.

[0057] The player information storage unit 222 stores information related to each player 3 who uses the game service. For example, the player information storage unit 222 stores the player ID ( 3) and the player name, game server name, and player ID linked to the player ID. Information such as the date of registration to the service, the player level, etc. may also be stored. The information storage unit 222 stores play data that is accumulated by each player 3 using the game service. The history information may be stored in association with the player ID. The play history information may include: For example, login date and time, number of logins, total play time, number of games played, Data when a game unit is cleared, items and characters owned by Player 3, etc. Game media, amount of paid or free in-game currency owned, amount charged, number of charges, and lottery results Data such as row counts may also be included.

[0058] The model storage unit 223 stores a predecessor model and a learned model with settings used for controlling the NPC. The model storage unit 223 stores the trained model together with the trained model. The learning data and the trained data to be used for constructing and relearning the rule may be stored.

[0059] <Server communication section> The server communication unit 23 communicates with the terminal communication unit 35 of the player terminal 30 via the network N. The server communication unit 23 and the terminal communication unit 35 of the player terminal 30 are connected to each other. The server 20 transmits and receives information from the player terminal 30 via the server communication unit 23. Receive various commands (prompts, requests, etc.) sent from the player terminal 30. Based on the various commands received, the control unit 21 executes processing according to the game program. The server communication unit 23 receives the processing result ( For example, game images, game sounds, etc.) are transmitted to the player terminal 30. A part of the program may be transmitted to the player terminal 30 and executed on the player terminal 30. good.

[0060] <Player device> FIG. 3 is a functional block diagram of the player terminal 30 shown in FIG. An information processing device that the player 3 can own and use (e.g., a smartphone, a mobile phone, (phones, tablets, game consoles, other computers, etc.) The terminal 30 transmits the processing result of the game program transmitted from the server 20 as a game application. Output via a web browser or application (e.g., game image output, game sound output, etc.) The player terminal 30 can, for example, control the entire player terminal 30. A control unit 31, a storage unit 32 for storing various data and programs, and a player inputting operations a display unit 34 for displaying a game screen, an operation screen, etc.; and a server 2. 0 and a terminal communication unit 35 for performing information communication with the

[0061] <Control unit (player terminal)> The control unit 31 controls each part of the player terminal 30, such as the display and the camera. The control unit 31 has the necessary functions, and each function is controlled by a CPU (Central Processor). ssing Unit) and GPU (Graphics Processing Unit) ) is realized by executing a program stored in a specified memory. The control unit 31 of the terminal 30 includes, for example, an operation information receiving unit 311, a game operation unit 312, and a processing unit. and a processing result receiving unit 313.

[0062] The operation information receiving unit 311 receives game information input by the player 3 to the input unit 33 (to be described later). The operation information receiving unit 311 receives the operation information. The game controller 313 executes a process of outputting operation information relating to the game to the game operation unit 312.

[0063] The game operation unit 312 executes processes for operating a game application. The game operation unit 312 executes the game operation information stored in the storage unit 22 described later. The software and game-related operations input by the player through the input unit 23 described below The game operation unit 312 starts and operates the game based on the information. An image for a game is generated from image data included in the game operation information, and the generated image is In the same manner, the game operation unit 312 performs a control process for outputting the game data to the display unit 34. The game's sound is generated from the music data and voice data included in the game operation information in accordance with the game's operation. A control process for generating music and sound and outputting the generated music and sound from the player terminal 30 Carry out the analysis.

[0064] Note that here, the native app executes some of the functions of the web app, i.e. The game operation unit 312 of the player terminal 30 has the functions of the processing result generation unit 215 of the server 20. The control unit 31 of the player terminal 30 is a game operation unit 31. In that case, the processing result generating unit 215 of the server 20 may not have the function of the processing As a result, all images and sounds for the game may be generated. Instead of the processing result generating unit 215 of the server 20, the game operation unit 312 of the player terminal 30 The information processing system 10 may be configured to generate all images and sounds for the game.

[0065] When the control unit 31 has a game operation unit 312, the game that the game operation unit 312 operates is The predetermined parameters in the game are managed by the server 20. For example, Parameters such as the number of players and in-game currency are managed by the server 20. The game operation unit 312 of the ear terminal 30 controls the change of these predetermined parameters in the game. When a process involving the above occurs, the server 20 communicates with the server 20 to manage the process. The game operation unit 312 then updates the parameters that it manages. The updated parameters are received from the server 20 and the game operation continues based on the updated parameters.

[0066] When the server 20 has a processing result generating unit 215, the server 20 receives the processing result of the player terminal 30. The receiving unit 313 receives the processing result generated and transmitted by the processing result generating unit 215 of the server 20. The player terminal 3 receives the information and outputs the processing result to the display unit 34. The control unit 31 of FIG. 0 includes the above-mentioned operation information receiving unit 311, game operation unit 312, and processing result unit 313. The result receiving unit 313 may have a functional unit other than the result receiving unit 313.

[0067] <Storage unit (player terminal)> The memory unit 32 of the player terminal 30 is used to execute various control processes and various functions in the control unit 31. It stores programs, input data, etc. for the ss Memory), ROM (Read Only Memory), SRAM (St atic Random Access Memory), flash memory, HDD ( Hard Disk Drive), SSD (Solid State Drive), Other storage, etc., or any combination of them. In addition, when the player terminal 30 performs communication, the storage unit 32 stores the communication status of the player terminal 30. The storage unit 32 may include a player setting storage unit 321.

[0068] The player setting storage unit 321 stores information about the player, such as information about the payment method and account information. Such player setting information is stored in the player information storage unit of the server 20. The number of times may be stored in the memory 222.

[0069] <Input section> The input unit 33 receives an input operation by the player 3 and outputs the player 3's The operation information is generated. The input operation by the player 3 may be, for example, an input operation using a touch panel or the like. Input operations by touching the display screen, or via physical buttons provided on the player terminal 30 Input operation, text input operation using a keyboard, voice input operation using a microphone, The player 3 plays the game using a sensor such as a gyro sensor installed in the layer terminal 30. Examples of input operations include an input operation by physically moving the ear terminal 30 (tilting, shaking, etc.). do.

[0070] When the player 3 talks to the NPC in the game, the player 3 touches the display screen, Alternatively, input operations may be performed using a keyboard, in which case text data is input as an operation. Alternatively, the player 3 may perform an input operation using a microphone. In this case, the voice data or the text data converted from the voice data is often manipulated. In the conversation between Player 3 and NPC, the above operation information is generated. is used as input information (prompts) to trained models such as trained models with settings. This may also be the case.

[0071] <Display section> The display unit 34 is a display that displays the results of the processing of the game program. The display is a touch panel that can accept input operations from the player 3 by touching the display surface. It's fine.

[0072] <Device communication unit> The terminal communication unit 35 is connected to the server communication unit 23 of the server 20 via the network N. The terminal communication unit 35 and the server communication unit 23 transmit and receive information to and from each other.

[0073] Next, the outline of the process of adding information to the character AI will be described with reference to the flowchart shown in FIG. 5A to FIG. 5B. This will be explained based on the above.

[0074] FIG. 5A shows a flowchart of a process for performing additional learning in response to acquisition of a first game medium by a player 3. As shown in FIG. 5A, in step S1, the player 3 selects any When the player acquires a game medium through a method (such as gacha or an event reward), the additional information acquisition unit 212 The game media provided to the player 3 includes a first game medium having setting information. It is determined whether or not the acquisition of the first game medium has been detected (step S2). If so, the process proceeds to the next step S3.

[0075] In step S3, the additional information acquisition unit 212 acquires the first game medium acquired by the player 3. The setting information associated with the body is acquired, and the acquired setting information is transmitted to the model generation unit 213. In step S4, the model generation unit 213 adds the setting information acquired by the additional information acquisition unit 212 to the model generation unit 213. A pre-trained model can be trained using any method, such as parallel training, fine-tuning, or transfer learning. At this time, the model generation unit 213 performs additional learning on the first game medium before the additional learning. Additional learning data including the associated setting information may be generated. Through additive learning, a trained model with settings is generated.

[0076] In step S4, the predecessor model used to control the NPC before the additional learning is may be left in the model storage unit 223 to generate a trained model with settings, or the predecessor model The model may be updated to generate a trained model with configuration.

[0077] In FIG. 5A, two examples of the flow of the information addition process are shown. After S4, the process may proceed to step S7 (example 1), or steps S5 and S6 may be performed. In the case of Example 1, the setting learning is performed in Step S4, and then the process proceeds to Step S7 (Example 2). After generating the trained model, the process proceeds to step S7 without going through steps S5 and S6 described below. Alternatively, the generated trained model with settings may be used to control the NPC. As shown below, after the generation of the learned model with settings, the first game medium of the player 3 is The processing in steps S5 and S6 is carried out so that the control of the NPC is switched upon use. The theory may be carried out.

[0078] When the predecessor model and the trained model with settings are made to coexist in a valid state, The control model may be switched by detecting the use of a first game medium. When player 3 selects the game medium to be used in the game during the preparation process before the game is executed, The PC control unit 211 selects the game medium selected by the player 3 as the game medium to be used. Specifically, whether or not a first game medium having additionally learned setting information is included (i.e. , whether the first game medium has been used or not (step S5). When the use of the first game medium is detected, the control model is learned from the predecessor model with settings. In step S6, the control model is switched to another model. The model shown is a replacement of the first model, but the trained model with settings is used without going through the procedure of switching models. Detects the use of a game medium and stores setting information associated with the first game medium in a NP in the game. By reflecting this in the activity of C, a process equivalent to switching the control model may be executed.

[0079] In step S7, the learned model with settings is used based on the function of the NPC control unit 211. The system then controls the NPCs based on the learned model. The control may be based only on the trained model with settings, or the trained model with settings and The control may be performed in conjunction with other trained models (such as predecessor models).

[0080] In a conversation between the player 3 and the NPC, the conversation (question) on the player 3's side is input to the input unit 33. Input text data and / or audio data to a pre-trained model with settings The trained model with the settings is then used to determine the NPC's response to the conversation on Player 3's side. The answer is determined and the decision is output as control information for the NPC. Since the setting information linked to the first game medium as shown in Figure 4 is reflected, It is possible to create new interest in the conversation that is different from the conversation before the additional learning.

[0081] For example, there are stories related to items such as the item “Pairing” in Figure 4. When incorporated into a trained model with settings as additional information, the item "pairing" is used By using this to play the game, conversations will unfold, including reminiscing about dates. In the case of a "diary," by using the item to play the game, lost memories will be restored. The NPC's conversation will be conducted in the revived setting. By having the PC wear it, the NPC will be able to behave (speech and / or actions) like a maid. In addition, in the case of the item "T-shirt with takoyaki design", By equipping an NPC with this system, the NPC will have a conversation in Osaka dialect. As shown, a new conversation is created that did not take place before the acquisition or use of the first game medium. This can increase the interest of the player 3 in a game involving conversations between NPCs. In the above example, the main control target is "conversation generation." It may also be used for image generation.

[0082] FIG. 5B shows a case where additional learning is performed in response to the use of the first game medium by the player 3. As shown in FIG. 5B, in step S1a, the player 3 When the player 3 selects the game content to be used in the game, the additional information acquisition unit 212 The game media selected as the game media for use include a first game medium having setting information. It is determined whether or not the first game medium is being used (step S2a). If so, the process proceeds to step S3. 7 is the same as in FIG. 5A, so the description will be omitted.

[0083] Figure 5C shows the flow of performing additional learning when an update to the game system is made. As shown in FIG. 5C, in step S1b, the game is updated. When a new game medium is additionally registered in the game information storage unit 221, the additional information acquisition unit 212 Whether or not the newly registered game media includes a first game media having setting information Based on this determination, it is determined whether the new first game medium can be updated (step S2b). If it is detected that the information has been added, the process proceeds to the next step S3.

[0084] Steps 3 and S4 in FIG. 5C are the same as those in FIG. 5A, so the explanation will be omitted. Even when additional learning is performed due to a request, the process proceeds to step S7 after step S4. Also, in FIG. 5C, like FIG. 5A, two flows are shown, and steps After S4, the process may proceed to step S7 via step S5b and step S6, or After step S4, steps S5b, S5c, and S6 are performed. You may then proceed to step S7.

[0085] When additional learning is performed after an update, the first game by player 3 is A process of switching the control model may be executed when the medium is acquired or used. For example, As shown in step S5b of FIG. 5C, after the trained model with the setting is generated, the NPC control unit 211, player 3 acquires the first game medium that was newly added through an update. Detects when the NPC is in a state where it is unable to control itself, and changes the control of the NPC from the predecessor model to the pre-trained model with settings. In step S5b, the player may execute a process of switching to the After the player 3 acquires the first game medium, in step S5c, the NPC control unit 211 Detects the use of the first game medium by Layer 3 and sets the control of the NPC from the predecessor model Alternatively, a process of switching to control that utilizes the trained model may be executed (step S6).

[0086] In step S6 of FIG. 5C, the control model is switched. Without following any procedure, the configured trained model detects the acquisition or use of the first game medium, By reflecting the setting information associated with the first game medium in the activities of NPCs in the game, A process equivalent to switching of the control model may be executed.

[0087] <Summary of the embodiment> The information processing system 10 of this embodiment is a game that uses non-player characters (NPCs). The information processing system 10 is a game system that provides at least one Based on the learned model, an NPC control unit 211 controls the non-player characters, an additional information acquisition unit 212 that acquires setting information associated with the game medium (first game medium); Based on the setting information associated with the game medium acquired by the information acquisition unit 212, a learned model with settings is The NPC control unit 211 includes a model generation unit 213 that generates a model. Based on the model, non-player characters are controlled by reflecting the setting information linked to the game media. To control.

[0088] By providing the information processing system 10 with the above features, it is possible to The activities of the game can be diversified according to the ownership status of the game media of the player 3. As a result, the interest of the player 3 in the game using the non-player characters is increased. It can be done.

[0089] In the information processing system 10, the additional information acquisition unit 212 acquires game media by the player 3. When acquisition or use of the game content is detected, setting information associated with the game content may be acquired. In addition, when the game system is updated, the additional information acquisition unit 212 acquires the game media information. Configuration information associated with the body may also be acquired.

[0090] In the information processing system 10, the NPC control unit 211 receives game media from the player 3. When the acquisition or use of the game media is detected, the setting information associated with the game media is This may be reflected in the activities of the player character.

[0091] As described above, when the player 3 acquires the game medium, the player 3 may perform additional learning or By changing the control, it is expected that players will be more motivated to acquire the game media. 3. The selection of the game media to be used as a trigger for additional learning or control changes By doing so, the player 3 can select the game media to be used based on his / her own will. Based on this, it is possible to create a variety of developments in games that use non-player characters. .

[0092] The configuration-trained model is a set of constraints related to the actions of non-player characters in a game. The model may also output control information, particularly related to the conversation of non-player characters. The model may output control information.

[0093] The above-described embodiment is merely an example for facilitating understanding of the present invention and does not limit the present invention. The present invention is not intended to be construed as being limited to the above-mentioned embodiments, and may be modified or improved without departing from the spirit and scope of the present invention. It goes without saying that the present invention includes equivalents thereof.

[0094] The present disclosure has the following configuration. (Item 1) An information processing system for providing a game using a non-player character, comprising: N that controls the non-player characters based on at least one learned model. A PC control unit, an additional information acquisition unit that acquires setting information associated with a game medium; Based on the setting information associated with the game medium acquired by the additional information acquisition unit, A model generation unit that generates a trained model, The NPC control unit performs the above-mentioned processing based on the learned model with the setting, while reflecting the setting information. An information processing system that controls a non-player character. (Item 2) The learned model with settings outputs control information related to the conversation of non-player characters. 2. The information processing system according to item 1, which is a model for outputting the information processing system. (Item 3) The additional information acquisition unit detects the acquisition or use of the game content by a player, and 3. The information processing system according to item 1 or 2, wherein the setting information is associated with a program medium. (Item 4) Further comprising a storage unit for storing game information; The additional information acquisition unit, when the game content is registered in the storage unit, 3. The information processing system according to item 1 or 2, which acquires the setting information associated with the (Item 5) The NPC control unit detects the acquisition or use of the game medium by a player, and Item 1, in which the fixed information is reflected in the activity of the non-player character in the game; or 2. An information processing system according to claim 2. (Item 6) Executing information processing related to the provision of games using non-player characters on a computer A program for causing Controlling the non-player character based on at least one learned model. And, Obtaining setting information associated with a game medium; A trained model with settings is generated based on the setting information associated with the acquired game content. To do, The setting information is used to control the non-player character based on the learned model with the setting. and a program causing the computer to execute the steps of: (Item 7) A computer executes information processing related to the provision of a game using non-player characters. An information processing method comprising: Controlling the non-player character based on at least one learned model. And, Obtaining setting information associated with a game medium; A trained model with settings is generated based on the setting information associated with the acquired game content. To do, The setting information is used to control the non-player character based on the learned model with the setting. and reflecting the result in the information processing method executed by the computer. [Explanation of symbols]

[0095] 1 … Business operator 3 … Player 10. Game System 20 ... Server (information processing device) 21 ... Control section 210 ... Game execution section 211 … NPC control unit 212 … Additional information acquisition department 213 ... Model generation section 214 … Operation information reception unit (server side) 215 ... Processing result generation unit 22 … Storage section 221 ... Game information storage unit 222 ... Player information storage unit 223 … Model memory section 23 ... Server communication section 30 … Player terminal 31 ... Control unit (player terminal side) 311 ... Operation information reception unit (player terminal side) 312 ... Game operation section 313 ... Processing result receiving unit 32 ... Memory unit (player terminal side) 321 ... Player setting memory section 33 … Input section 34 … Display section 35 … Terminal communication section

Claims

1. An information processing system for providing a game using a non-player character, comprising: N, which controls the non-player character based on at least one learned model. A PC control unit; an additional information acquisition unit that acquires setting information associated with a game medium; Based on the setting information associated with the game medium acquired by the additional information acquisition unit, A model generation unit that generates a trained model, The NPC control unit performs the above-mentioned processing based on the learned model with the setting, while reflecting the setting information. An information processing system that controls a non-player character.

2. The learned model with settings outputs control information related to the conversation of non-player characters.

2. The information processing system according to claim 1, wherein the model is a

3. The additional information acquisition unit detects the acquisition or use of the game content by a player, and The information processing system according to claim 1 or 2, wherein the setting information is associated with a program medium.

4. Further comprising a storage unit for storing game information; The additional information acquisition unit, when the game content is registered in the storage unit, The information processing system according to claim 1 or 2, wherein the setting information associated with the first and second information is acquired.

5. The NPC control unit detects the acquisition or use of the game medium by a player, The information is reflected in the activities of the non-player characters in the game.

2. An information processing system according to claim 2.

6. Executing information processing related to the provision of games using non-player characters on a computer A program for causing Controlling the non-player character based on at least one learned model. And, Obtaining setting information associated with a game medium; A trained model with settings is generated based on the setting information associated with the acquired game content. To do, The setting information is used to control the non-player character based on the learned model with the setting. and a program causing the computer to execute the steps of:

7. A computer executes information processing related to the provision of a game using non-player characters. An information processing method comprising: Controlling the non-player character based on at least one learned model. And, Obtaining setting information associated with a game medium; A trained model with settings is generated based on the setting information associated with the acquired game content. To do, The setting information is used to control the non-player character based on the learned model with the setting. and reflecting the result in the information processing method executed by the computer.

Images