Game interaction simulation method and device, computer device and storage medium

Abstract

The application provides a game interaction simulation method and device, computer equipment and a storage medium. The method comprises the following steps: obtaining game confrontation data in a target game session, wherein the target game session at least comprises a user-controlled role and a man-machine-controlled role; determining confrontation label information of the man-machine-controlled role corresponding to the game confrontation data based on an analysis result of the game confrontation data; and controlling the man-machine-controlled role to output interaction information associated with the confrontation label information to the user-controlled role. The application associates the game confrontation data with the interaction information by using the confrontation label information, can more conveniently train a language model, can better process the interaction information, thereby enriching the optional interaction information of the game event and improving the personification effect of the AI.

Description

Technical Field

[0001] This application relates to the field of computer technology, and specifically to a game interaction simulation method, apparatus, computer device, and storage medium (computer-readable storage medium). Background Technology

[0002] In many games, the player experience is a key factor in improving user retention. Take MOBA (Multiplayer Online Battle Arena) games as an example: players on the same team compete against players on other teams through tactical cooperation to achieve a final objective and win. Therefore, winning in MOBA games enhances the player experience. In some cases, AI-controlled game characters need to interact and engage with player-controlled characters.

[0003] However, current AI interaction information in games is often just a pre-set, fixed type of content. That is, existing game interaction methods typically generate specific interaction information for the AI ​​character after a specific game event is triggered. Therefore, to improve the human-like effect of AI interaction, it is often necessary to assign richer interaction information to the same game event. However, manually configuring interaction information is cumbersome, and interaction information extracted from other game matches is often difficult to label with game events for model training. This makes it difficult for existing game interaction methods to configure rich interaction information for game events, thus affecting the human-like effect of the AI. Summary of the Invention

[0004] Therefore, it is necessary to provide a game interaction simulation method, device, computer equipment, and storage medium to address the aforementioned technical problems and solve the technical issue of poor AI anthropomorphism in existing games.

[0005] Firstly, this application provides a game interaction simulation method, including:

[0006] Acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character;

[0007] Based on the analysis results of the game confrontation data, the confrontation tag information of the human-computer controlled character corresponding to the game confrontation data is determined;

[0008] The human-machine control role is controlled to output interactive information associated with the adversarial tag information to the user control role.

[0009] Secondly, this application provides a game interaction simulation device, comprising:

[0010] The acquisition module is used to acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character;

[0011] The parsing module is used to determine the adversarial tag information of the human-computer controlled character corresponding to the game adversarial data based on the parsing results of the game adversarial data;

[0012] The output module is used to control the human-machine control role to output interactive information associated with the adversarial tag information to the user control role.

[0013] Thirdly, this application also provides a computer device, characterized in that the computer device comprises:

[0014] One or more processors;

[0015] Memory; and

[0016] One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the game interaction simulation method provided above.

[0017] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, the computer program being loaded by a processor to execute the game interaction simulation method provided above.

[0018] Fifthly, embodiments of this application provide a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the game interaction simulation method described above.

[0019] The game interaction simulation method provided in this application collects game confrontation data in real time during game matches involving at least one user-controlled character and one AI-controlled character. Then, based on the analysis results of the game confrontation data, it determines the confrontation label information corresponding to the game confrontation data for the AI-controlled character. Finally, it controls the AI-controlled character to output interactive information associated with the confrontation label information to the user-controlled character. By using the confrontation label information to indirectly link game confrontation data and interactive information, compared to the prior art which directly selects corresponding interactive information based on game events in the game confrontation data, reasonable confrontation label information can more easily train and generate a language model, and can better process interactive information in other game matches, thereby enriching the optional interactive information of game events and improving the anthropomorphic effect of AI. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This application provides a flowchart illustrating the steps of a game interaction simulation method.

[0022] Figure 2 This application provides a flowchart illustrating the steps for determining adversarial tag information based on game data parsing in an embodiment of the present application.

[0023] Figure 3 This application provides a schematic flowchart illustrating the steps of associating interactive information with tag information and storing it in a database.

[0024] Figure 4 A schematic diagram illustrating the steps of training a language classification model based on a semi-supervised training method, as provided in an embodiment of this application;

[0025] Figure 5 This application provides a flowchart illustrating the steps for outputting interactive information in an embodiment.

[0026] Figure 6 This application provides a flowchart illustrating the steps for outputting interactive text information in an embodiment.

[0027] Figure 7 A flowchart illustrating the steps of a game interaction simulation method based on interaction frequency, provided in this application embodiment;

[0028] Figure 8 This is a schematic diagram of the structure of a game interaction simulation device provided in an embodiment of this application;

[0029] Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0030] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0032] In the description of this application, the word "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0033] To facilitate understanding of the game interaction simulation method provided in this application embodiment, the application scenarios of the game interaction simulation method are first described. The game interaction simulation method provided in this application embodiment is mainly applied to game matches involving human-computer controlled characters, i.e., AI-controlled characters. Specifically, taking MOBA games as an example, by matching players into the same team and providing an AI-controlled character in the other team to facilitate player victory, the user's gaming experience is improved. In the above process, appropriate interactions initiated by the human-computer controlled character can increase the anthropomorphic effect of the human-computer controlled character; conversely, unreasonable interactions initiated by the human-computer controlled character result in a poor anthropomorphic effect. Therefore, improving the anthropomorphic effect of the human-computer controlled character, that is, controlling the human-computer controlled character to initiate reasonable interactions in appropriate game scenarios, can significantly affect the user's gaming experience.

[0034] Currently, the interactive information initiated by AI-controlled characters in games is often pre-set and fixed; that is, after triggering a specific game event, the AI-controlled character will output specific interactive information. However, repeatedly outputting the same interactive information is not a reasonable interactive behavior. Therefore, to improve the anthropomorphic effect of the AI-controlled character, it is often necessary to configure more interactive information for the same game event for the AI-controlled character to choose from, which means establishing a mapping relationship between interactive information and game events. Manually configuring the mapping relationship between interactive information and game events is costly and time-consuming. Therefore, the current approach is mainly to establish the mapping relationship between interactive information and game events by training a network model. This involves collecting interactive information from historical game matches, manually labeling it with corresponding game event tags, and training a neural network model that describes the mapping relationship between different interactive information and game events. Subsequently, this neural network model can be used to classify other interactive information and associate them with different game events. Thus, during gameplay, after triggering a specific game event, the AI-controlled character can output richer interactive information, improving the anthropomorphic effect of the AI-controlled character.

[0035] However, in practice, it has been found that due to the diversity of game events, if game events are chosen as labels, training a network model that describes the relationship between interaction information and game events is often difficult to implement, often resulting in insufficient accuracy of the trained network model, thus affecting the final game interaction process. To address this problem, this application provides a game interaction simulation method. This method is typically installed as a computer program on a game interaction device, which is usually a processor located in a computer device. The game interaction device in the computer device executes the game interaction simulation method provided in this application by running the computer program corresponding to the game interaction method. Specifically, as... Figure 1 As shown, Figure 1 This application provides a flowchart illustrating the steps of a game interaction simulation method, specifically including steps S110 to S130:

[0036] S110, acquire game confrontation data in the target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character.

[0037] Based on the application scenarios of the game interaction simulation method provided above, it can be seen that the game interaction simulation method provided in this application embodiment is mainly applied to game matches with human-computer controlled characters. Therefore, the target game match includes at least one user-controlled character and one human-computer controlled character. Taking the aforementioned MOBA game as an example, the user-controlled character is the character controlled by the player, while the human-computer controlled character usually refers to the character controlled by artificial intelligence to compete against the user-controlled character.

[0038] Building upon the aforementioned foundation, the game interaction simulation device collects real-time game data during gameplay. Specifically, this data primarily includes movement, attack, and skill activation commands issued by the user to their controlled character, as well as movement, attack, and skill activation commands issued by the AI ​​to the AI-controlled character, and information on character attribute changes resulting from these commands, such as changes in character health. Of course, the above data is merely one example; in fact, any information generated during a game that can be used to analyze the real-time battle situation can be considered game data.

[0039] S120, based on the analysis results of the game confrontation data, determine the confrontation tag information of the human-computer control character corresponding to the game confrontation data.

[0040] In this embodiment, by analyzing game-related data, game events can typically be filtered out, thereby further determining the adversarial tag information corresponding to the human-computer controlled character's game-related data. Specifically, unlike game events, the adversarial tag information here is not game events like "kill," "team fight," "resource grab," or "tower steal," but rather tag information such as "praise," "taunt," "dissatisfaction," and "blame," filtered based on user sentiment analysis. This tag information is usually generated based on game events such as "kill," "team fight," "resource grab," and "tower steal." For example, when a kill is achieved or a team fight is won, the adversarial tag information corresponding to the human-computer controlled character is often praising one's own team or taunting the enemy; conversely, when killed or a team fight is lost, the adversarial tag information corresponding to the human-computer controlled character is often praising the opponent or blaming teammates. Specifically, as... Figure 2 As shown, Figure 2 This application provides a flowchart illustrating the steps for determining adversarial tag information based on game data parsing, specifically including steps S210 to S220:

[0041] S210, based on the parsing results of the character data corresponding to the human-computer controlled character and the character data corresponding to the user-controlled character in the game confrontation data, determine the game event between the human-computer controlled character and the user-controlled character.

[0042] As can be seen from the foregoing description, by analyzing the character data corresponding to the AI-controlled character and the user-controlled character in the game's competitive data, game events between the AI-controlled character and the user-controlled character can usually be filtered out. For example, taking two relatively simple kill methods, "tower dive kill" and "low-health comeback kill," by analyzing the character data of character A and character B in the game's combat data, it can be seen that if character B's position data indicates that character B is within a specific range, that is, under a turret, and character A kills character B by reasonably moving, attacking, and releasing skill commands, that is, reducing character B's health attribute to below zero, then the "tower dive kill" event can be considered triggered. Similarly, if by analyzing the character data of character A and character B in the game's combat data, it can be seen that if character A, with lower health attribute than character B, avoids the damage caused by character B's attacks and skill commands by reasonably moving, attacking, and releasing skill commands, and kills character B, then the "low-health comeback kill" event can be considered triggered. Of course, the above example is only one feasible event judgment logic. The judgment logic of other game events, including but not limited to the aforementioned "team fight," "resource grab," "tower steal," etc., will not be elaborated here in this application embodiment.

[0043] S220, based on the preset association between events and emotion tags, the emotion tags associated with the game events are set as the adversarial tag information of the game adversarial data corresponding to the human-computer control character.

[0044] In this embodiment, the emotion tags are the aforementioned tags such as "praise (teammate or opponent)," "taunt (opponent)," "dissatisfaction (teammate)," and "blame (teammate)." After determining the game event between the AI-controlled character and the user-controlled character, such as "kill," based on the pre-set association, the emotion tag associated with this game event can be set as the corresponding adversarial tag information for the AI-controlled character. For example, when the AI-controlled character triggers the game event of "being killed while low on health," the emotion tag associated with this event is "praise the opponent," and the corresponding adversarial tag information for the AI-controlled character is "praise the opponent." Or, when the AI-controlled character triggers the game event of "being tower-dived and killed by the opponent without teammate support," the emotion tag associated with this event is "dissatisfaction with teammates not supporting," and the corresponding adversarial tag information for the AI-controlled character is "dissatisfaction with teammates not supporting."

[0045] It should be noted that the association between game events and emotion tags can be pre-configured manually, but configuring emotion tags corresponding to game events is more convenient than configuring interaction information. Furthermore, the above explanation uses a single emotion tag for a game event as an example. In reality, a game event may correspond to multiple emotion tags; for example, "team battle loss" could correspond to both "praising the opponent" and "blaming teammates." Regardless, the above will not affect the implementation of the embodiments of this application.

[0046] S130, control the human-machine control role to output interactive information associated with the adversarial tag information to the user control role.

[0047] After determining the adversarial tag information corresponding to the human-computer controlled character, the game interaction simulation device can further control the human-computer controlled character to output interactive information associated with the adversarial tag information to the user-controlled character. It should be noted that outputting interactive information to the user-controlled character essentially refers to outputting interactive information to the player user corresponding to the user-controlled character. For example, when the adversarial tag information is "praise the opponent," the interactive information associated with this adversarial tag information includes text interactive information such as "show off" or "plays really well," as well as emoticon interactive information such as "thumbs up." The game interaction simulation device can then control the human-computer controlled character to output any one or any combination of the above interactive information to the user-controlled character.

[0048] Furthermore, to enhance the anthropomorphic effect of game interaction, the same adversarial tag information typically corresponds to multiple interaction messages. These interaction messages are often extracted in advance from a large number of game matches and pre-associated with the adversarial tag information, stored in a database. This allows the game interaction simulation device to directly query the database and retrieve the corresponding interaction information after determining the adversarial tag information of the human-computer controlled character. Specifically, for example... Figure 3 As shown, Figure 3 This application provides a flowchart illustrating a step of associating and storing interactive information and tag information in a database, specifically including steps S310 to S330:

[0049] S310 extracts historical interaction information from historical sample games.

[0050] In this embodiment of the application, historical sample games refer to game games between other players recorded by the game system. The historical interaction information in the historical sample games is the real interaction information that has occurred between players. Therefore, using this historical interaction information as the interaction information for candidate human-computer control characters can further improve the anthropomorphic effect of human-computer control characters.

[0051] Of course, in the process of extracting historical interaction information, in order to avoid the bad gaming experience brought to players by bad interaction information, the game interaction simulation device will further clean the historical interaction information and remove the bad interaction information in the historical interaction information. Of course, the preprocessing process of historical interaction information will not be described in detail in this embodiment of the application. For example, a common method is to filter out historical interaction information by setting "filter words".

[0052] Furthermore, in addition to considering negative interaction information in the interaction information, the historical interaction information extracted from historical sample games will also fully consider the frequency of occurrence of each historical interaction information. Under normal circumstances, the game interaction simulation device will prioritize extracting historical interaction information with higher occurrence frequency for subsequent processing.

[0053] S320, The historical interaction information is processed based on the pre-trained language classification model to obtain the label information corresponding to the historical interaction information.

[0054] In this embodiment, after obtaining the historical interaction information, the historical interaction information is further processed by the trained language classification model to obtain the label information corresponding to the historical interaction information. Specifically, the label information here is the emotional label provided above, such as "praise (teammate or opponent)", "taunt (opponent)", "dissatisfaction (teammate)", "blame (teammate)". Compared with using game events as labels, the language classification model is easier to train. For example, based on a certain amount of samples, a language classification model that can classify "played really well" as "praise (opponent)" can be effectively trained, but it is difficult to train a language classification model that can classify "played really well" as game event labels such as "killed" or "team fight lost".

[0055] Specifically, the language classification model here can be trained based on a neural network model. For example, a common method is supervised training. However, considering the large amount of interaction information data, sample labeling requires significant human resources. Therefore, as a feasible embodiment of this application, such as... Figure 4 As shown, Figure 4 This application provides a flowchart illustrating the steps for training a language classification model using a semi-supervised training method, specifically including steps S410 to S440:

[0056] S410, acquire the first sample interaction information and the second sample interaction information carrying the sample label.

[0057] In this embodiment of the application, the first sample interaction information refers to the interaction information that has not been processed and does not carry sample tags, while the second sample interaction information refers to the interaction information that has been preprocessed based on manual identification and carries sample tags. In general, the number of first sample interaction information is greater than the number of second sample interaction information.

[0058] S420, an initial language classification model is trained based on the second sample interaction information and the sample labels corresponding to the second sample interaction information.

[0059] In this embodiment, an initial language classification model can be trained based on manually labeled sample interaction information and the concept of supervised machine learning. Specifically, the model architecture of the initial language classification model can be set according to actual needs. For example, as a feasible embodiment of this application, the language classification model can use a pre-trained BERT (Bidirectional Encoder Representations from Transformers) model, that is, the encoder part of the standard Transformer structure is used as the main basic part. The specific model structure of the language classification model will not be described in detail here.

[0060] S430, based on the initial language classification model, process the first sample interaction information and output the predicted label information corresponding to the first sample interaction information.

[0061] In this embodiment of the application, after training an initial language classification model based on manually labeled sample interaction information, the predicted label information corresponding to the first sample interaction information that does not carry sample labels can be fitted to a certain extent based on the initial language classification model.

[0062] S440, based on the first sample interaction information and the predicted label information corresponding to the first sample interaction information, the second sample interaction information and the sample label corresponding to the second sample interaction information, the initial language classification model is updated to obtain the trained language classification model.

[0063] In this embodiment of the application, after generating the predicted label information corresponding to the first sample interaction information, the initial language classification model is further iteratively updated based on the first sample interaction information and its corresponding predicted label information, as well as the second sample interaction information and its corresponding sample label, which can further improve the classification effect of the initial language classification model until the final language classification model is obtained.

[0064] The language classification model based on semi-supervised training provided in this application does not rely on a large number of annotated sample interactive texts, and ensures the classification effect on other interactive texts in the future.

[0065] S330, the historical interaction information is associated with the tag information and stored in a preset database.

[0066] In this embodiment of the application, after obtaining the historical interaction information, the historical interaction information is associated with the tag information and stored in a preset database. Subsequently, after determining the adversarial tag information corresponding to the human-machine control role, the target interaction information associated with this adversarial tag information can be found directly by querying the database. The human-machine control role is then controlled to output the target interaction information to the user control role, thereby enabling the human-machine control role to quickly and accurately complete the anthropomorphic interaction of the human-machine control role.

[0067] The game interaction simulation method provided in this application collects game confrontation data in real time during game matches involving at least one user-controlled character and one AI-controlled character. Then, based on the analysis results of the game confrontation data, it determines the confrontation label information corresponding to the game confrontation data for the AI-controlled character. Finally, it controls the AI-controlled character to output interactive information associated with the confrontation label information to the user-controlled character. By using the confrontation label information to indirectly link game confrontation data and interactive information, compared to the prior art which directly selects corresponding interactive information based on game events in the game confrontation data, reasonable confrontation label information can more easily train and generate a language model, and can better process interactive information in other game matches, thereby enriching the optional interactive information of game events and improving the anthropomorphic effect of AI.

[0068] Furthermore, based on the aforementioned solutions, and considering that the interactive information associated with adversarial tag information typically includes multiple types, this application also provides an embodiment for outputting interactive information to the user control role in order to select appropriate interactive information for the human-machine control role. Figure 5 As shown, Figure 5 This application provides a flowchart illustrating the steps for outputting interactive information, specifically including steps S510 to S520:

[0069] S510, control the human-machine control role to output facial expression interaction information associated with the adversarial tag information to the user control role.

[0070] In this embodiment of the application, as can be seen from the foregoing description of the interactive information, the interactive information can generally include emoticon interactive information and text interactive information. For example, the most common emoticon interactive information is generally an interactive information displayed above the character's head to quickly show the player's mood. For example, common emoticon interactive information includes "crying", "thumbs up", and "angry" emoticon interactive information. Text interactive information is usually interactive information displayed in the chat bar in the form of text language. The specific content of emoticon interactive information and text interactive information will not be described in detail in this embodiment of the application.

[0071] Building upon the aforementioned, to facilitate better interaction between the human-computer control character and the player user who enjoys interaction, the game interaction simulation device will first control the human-computer control character to output facial expression interaction information associated with the confrontation tag information to the user-controlled character, in an attempt to request interaction with the player user.

[0072] S520, in response to a feedback operation instruction executed by the target user controlling the user control role, the human-machine control role is controlled to output text interaction information associated with the adversarial tag information to the user control role.

[0073] In this embodiment of the application, when the target user control character that controls the user control character executes the corresponding feedback operation instruction, such as the common practice of responding to the facial expression interaction information initiated by the human-computer control character, it can be assumed that the target user is a user player who likes interaction. At this time, the game interaction simulation device will further control the human-computer control character to output text interaction information associated with the confrontation tag information to the user control character. Otherwise, the game interaction simulation device will control the human-computer control character to stop the interaction.

[0074] Of course, when there are multiple text interaction messages, the game interaction simulation device will also fully consider other information when controlling the human-computer control character to output text interaction messages to the user control character. For example, as a feasible embodiment of this application, it can also consider the target user's feedback operation instructions, or consider the role type corresponding to the human-computer control character. Specifically, such as... Figure 6 As shown, Figure 6 This application provides a flowchart illustrating the steps for outputting interactive text information, specifically including steps S610 to S620:

[0075] S610, control the human-machine control role to output text interaction information associated with the adversarial tag information and the feedback operation command to the user control role.

[0076] In this embodiment, the control feedback operation instructions executed by the player user control character can reflect the player user's emotions to a certain extent. For example, if the player user's emotions are divided into positive and negative emotions, for negative emotions, the game interaction simulation device will prioritize the positive and affirmative text interaction information output by the human-computer control character to the user control character, while avoiding text interaction information with obvious negative attitudes such as sarcasm and blame, thereby improving the user's game experience.

[0077] S620, based on the role type corresponding to the human-machine control role, output the target text interaction information associated with the role type in the text interaction information to the user control role.

[0078] In this embodiment, the character type corresponding to the human-computer controlled character can, to some extent, describe the image of the player operating this human-computer controlled character. For example, when the character type corresponding to the human-computer controlled character is a support / assistant type, the game interaction simulation device will prioritize controlling the human-computer controlled character to output relatively mild-toned target text interaction information to the user-controlled character. Conversely, if the character type corresponding to the human-computer controlled character is an assassin / roaming type, the game interaction simulation device will prioritize controlling the human-computer controlled character to output more aggressive target text interaction information to the user-controlled character. Through the above processing, the output text interaction information can be made more consistent with the character image of the human-computer controlled character.

[0079] Of course, in addition to the above methods, the game interaction simulation device can also choose other reasonable methods to output corresponding text interaction information to the user-controlled character, which will not be described in detail here.

[0080] Furthermore, considering the timeliness of the interaction, as a feasible embodiment of this application, the time interval between the feedback operation command executed by the target user control character and the facial expression interaction information issued by the human-computer control character can be calculated. If the time exceeds a certain interval threshold, it can be considered that the feedback operation command and the facial expression interaction information are unrelated. Otherwise, it is considered that the feedback operation command is directed at the facial expression interaction information. At this time, the game interaction simulation device will select to control the human-computer control character to output text interaction information to the user control character.

[0081] Furthermore, before executing the game interaction simulation method provided in this application embodiment, as a feasible embodiment of this application, the game interaction simulation device can also obtain the historical game situation of the user corresponding to each user-controlled character in the target game game, and extract the interaction frequency of each user from the game. Since the interaction frequency can describe the user's activity level to a certain extent, it can be determined whether to execute the game interaction simulation method provided in this application embodiment based on the interaction frequency. Specifically, as shown in... Figure 7 As shown, Figure 7 A flowchart illustrating the steps of a game interaction simulation method based on interaction frequency, provided in this application embodiment, specifically includes steps S710 to S720:

[0082] S710 retrieves the historical game matches of the user corresponding to each user-controlled character in the target game match.

[0083] In this embodiment of the application, the user corresponding to the user control role is the player user, and the player user's historical game matches are usually recorded in the game system.

[0084] S720, if the interaction frequency of the user in the historical game matches meets the preset interaction conditions, then obtain the game confrontation data in the target game match.

[0085] Based on the foregoing, if the frequency of user interaction in historical game matches meets the preset interaction conditions, for example, when it exceeds a certain frequency threshold, the player user can be considered to be a target user with high activity and a preference for game interaction. In this case, the game interaction simulation will execute the game interaction simulation method provided in the embodiments of this application to improve the anthropomorphic effect of the human-computer control character.

[0086] To facilitate understanding of the complete implementation scheme of the game interaction simulation method provided in the embodiments of this application, the following will be combined with the foregoing Figures 1 to 7 The provided content offers a complete implementation process for a game interaction simulation method. Specifically, this game interaction simulation method mainly includes the following steps:

[0087] 1) Design the following tags: taunting / blaming teammates; expressing dissatisfaction with being targeted by the opponent; expressing dissatisfaction with teammates not helping when targeted; self-deprecatingly admitting poor performance; wanting to quit and surrender; taunting the opponent; praising the opponent; encouraging teammates; blaming a specific teammate (e.g., the jungler); praising oneself; and other categories, etc. Simultaneously, establish a mapping relationship between different game events and the above tags. For example, a game event where the AI-controlled character completes a kill as the killer can be associated with "taunting the opponent" and "praising oneself." Conversely, a game event where the AI-controlled character completes a kill as the victim can be associated with "praising the opponent" or "self-deprecatingly admitting poor performance." The specific game events and their mapping relationships with the tags can be based on actual game settings.

[0088] 2) Label some of the interaction information in the collected sample interaction information with the above labels, and train the pre-trained BERT model based on the labeled sample interaction information to obtain the initial language classification model.

[0089] 3) Use the initial language classification model to process the collected unlabeled sample information to obtain the predicted label information corresponding to each unlabeled sample information as its corresponding label information;

[0090] 4) The initial language classification model is further iterated and updated using the interaction information and its corresponding label information until the final trained language classification model is obtained;

[0091] 5) By preprocessing the historical interaction information in historical sample games with sensitive word filtering, bad interaction information is filtered out and the historical interaction information that appears frequently is retained.

[0092] 6) Use the trained language classification model to process the historical interaction information, obtain its corresponding classification label, and store the interaction information and classification label in a preset database.

[0093] 7) When at least one AI-controlled character is provided in a game match, retrieve the historical game matches of the user corresponding to each user-controlled character in that game match;

[0094] 8) If the frequency of user interaction in a historical game is greater than the preset frequency threshold, then activate game battlefield awareness to obtain game confrontation data in the target game in real time.

[0095] 9) Based on the analysis results of the character data corresponding to the human-computer controlled character and the character data corresponding to the user-controlled character, determine the game events between the human-computer controlled character and the user-controlled character;

[0096] 10) Utilize the aforementioned relationships between game events and tags to determine the corresponding adversarial tag information;

[0097] 11) Read interaction information from preset data based on adversarial tag information, including facial expression interaction information and text interaction information;

[0098] 12) Control the human-computer interaction character to output facial expression interaction information associated with the adversarial label information to the user control character;

[0099] 13) If a feedback operation instruction executed by the target user control role is obtained within a preset time interval, then appropriate target text interaction information is selected from the text interaction information and fed back based on the feedback operation instruction and / or the role type corresponding to the human-computer control role.

[0100] Through steps 1) to 6) above, a number of interaction information and their corresponding classification tags can be collected offline. Furthermore, the interaction information stored in the database can be continuously updated based on actual game data, and the above data is not affected by game type, thus having good portability.

[0101] Through steps 7) to 13) above, a specific implementation scheme for controlling the human-computer control role to output appropriate interactive information is provided, which can further improve the anthropomorphic interactive effect of the human-computer control role.

[0102] To better implement the game interaction simulation method provided in the embodiments of this application, based on the game interaction simulation method provided in the embodiments of this application, this application also provides a game interaction simulation device, such as... Figure 8 As shown, the game interaction simulation device 800 includes:

[0103] The acquisition module 810 is used to acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character.

[0104] The parsing module 820 is used to determine the adversarial tag information of the human-computer controlled character corresponding to the game adversarial data based on the parsing results of the game adversarial data;

[0105] The output module 830 is used to control the human-machine control role to output interactive information associated with the adversarial tag information to the user control role.

[0106] In some embodiments of this application, the output module 830 is used to extract historical interaction information from historical sample games; process the historical interaction information based on a pre-trained language classification model to obtain tag information corresponding to the historical interaction information; associate the historical interaction information with the tag information and store it in a preset database; query the preset database to obtain target interaction information associated with the adversarial tag information; and control the human-machine control role to output the target interaction information to the user control role.

[0107] In some embodiments of this application, the output module 830 is configured to acquire first sample interaction information and second sample interaction information carrying sample labels; train an initial language classification model based on the second sample interaction information and the sample labels corresponding to the second sample interaction information; process the first sample interaction information based on the initial language classification model to output predicted label information corresponding to the first sample interaction information; and update the initial language classification model according to the first sample interaction information and the predicted label information corresponding to the first sample interaction information, the second sample interaction information and the sample labels corresponding to the second sample interaction information, to obtain a trained language classification model.

[0108] In some embodiments of this application, the parsing module 820 is used to determine the game event between the human-computer controlled character and the user-controlled character based on the parsing results of the character data corresponding to the human-computer controlled character and the character data corresponding to the user-controlled character in the game confrontation data; and based on the preset association relationship between the event and the emotion tag, set the emotion tag associated with the game event as the confrontation tag information of the human-computer controlled character corresponding to the game confrontation data.

[0109] In some embodiments of this application, the output module 830 is used to control the human-machine control role to output facial expression interaction information associated with the adversarial tag information to the user control role; and in response to a feedback operation instruction executed by the target user controlling the user control role, control the human-machine control role to output text interaction information associated with the adversarial tag information to the user control role.

[0110] In some embodiments of this application, the output module 830 is used to control the human-machine control role to output text interaction information associated with the adversarial tag information and the feedback operation instructions to the user control role.

[0111] In some embodiments of this application, the output module 830 is used to output the target text interaction information associated with the role type in the text interaction information to the user control role based on the role type corresponding to the human-machine control role.

[0112] In some embodiments of this application, the output module 830 is used to obtain the feedback interval between the feedback operation instruction and the facial expression interaction information; if the feedback interval is less than a preset interval threshold, the human-computer control role is controlled to output text interaction information associated with the adversarial tag information to the user control role.

[0113] In some embodiments of this application, the acquisition module 810 is used to acquire the historical game matches of the user corresponding to each user-controlled character in the target game match; if the interaction frequency of the user in the historical game match meets the preset interaction conditions, then the game confrontation data in the target game match is acquired.

[0114] Specific limitations regarding the game interaction simulation device can be found in the limitations of the game interaction simulation method described above, and will not be repeated here. Each module in the aforementioned game interaction simulation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0115] In some embodiments of this application, the game interaction simulation device 800 can be implemented as a computer program, and the computer program can be implemented in, for example... Figure 9 The game runs on the computer device shown. The computer device's memory can store the various program modules that make up the game interaction simulation device 800, for example, Figure 8 The acquisition module 810, parsing module 820, and output module 830 are shown. The computer program comprised of these modules causes the processor to execute the steps in the game interaction simulation methods of the various embodiments of this application described in this specification.

[0116] For example, Figure 9 The computer device shown can be used as follows Figure 8The acquisition module 810 in the game interaction simulation device 800 shown executes step S110. The computer device can execute step S120 through the parsing module 820. The computer device can execute step S130 through the output module 830. The computer device includes a processor, memory, and network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with external computer devices via a network connection. When the computer program is executed by the processor, it implements a game interaction simulation method.

[0117] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0118] In some embodiments of this application, a computer device is provided, including one or more processors; memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to perform the following steps:

[0119] Acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character;

[0120] Based on the analysis results of the game confrontation data, the confrontation tag information of the human-computer controlled character corresponding to the game confrontation data is determined;

[0121] The human-machine control role is controlled to output interactive information associated with the adversarial tag information to the user control role.

[0122] In some embodiments of this application, when the processor executes the computer program, it further performs the following steps: extracting historical interaction information from historical sample games; processing the historical interaction information based on a pre-trained language classification model to obtain tag information corresponding to the historical interaction information; associating the historical interaction information with the tag information and storing it in a preset database; querying the preset database to obtain target interaction information associated with the adversarial tag information; and controlling the human-machine control role to output the target interaction information to the user control role.

[0123] In some embodiments of this application, when the processor executes the computer program, it further performs the following steps: acquiring first sample interaction information and second sample interaction information carrying sample labels; training an initial language classification model based on the second sample interaction information and the sample labels corresponding to the second sample interaction information; processing the first sample interaction information based on the initial language classification model and outputting predicted label information corresponding to the first sample interaction information; updating the initial language classification model according to the first sample interaction information and the predicted label information corresponding to the first sample interaction information, the second sample interaction information and the sample labels corresponding to the second sample interaction information, to obtain a trained language classification model.

[0124] In some embodiments of this application, when the processor executes the computer program, it further implements the following steps: based on the parsing results of the character data corresponding to the human-computer controlled character and the character data corresponding to the user-controlled character in the game confrontation data, it determines the game event between the human-computer controlled character and the user-controlled character; based on the preset association relationship between the event and the emotion tag, it sets the emotion tag associated with the game event as the confrontation tag information corresponding to the human-computer controlled character in the game confrontation data.

[0125] In some embodiments of this application, when the processor executes the computer program, it further implements the following steps: controlling the human-machine control role to output facial expression interaction information associated with the adversarial tag information to the user control role; and in response to a feedback operation instruction executed by the user control role under the control of the target user, controlling the human-machine control role to output text interaction information associated with the adversarial tag information to the user control role.

[0126] In some embodiments of this application, when the processor executes the computer program, it also performs the following steps: controlling the human-machine control role to output text interaction information associated with the adversarial tag information and the feedback operation instructions to the user control role.

[0127] In some embodiments of this application, when the processor executes the computer program, it also performs the following steps: outputting target text interaction information associated with the role type in the text interaction information to the user control role based on the role type corresponding to the human-machine control role.

[0128] In some embodiments of this application, when the processor executes the computer program, it further implements the following steps: obtaining the feedback interval between the feedback operation instruction and the facial expression interaction information; if the feedback interval is less than a preset interval threshold, controlling the human-computer control role to output text interaction information associated with the adversarial tag information to the user control role.

[0129] In some embodiments of this application, when the processor executes the computer program, it also performs the following steps: obtaining the historical game matches of the user corresponding to each user-controlled character in the target game match; if the interaction frequency of the user in the historical game match meets the preset interaction conditions, then obtaining the game confrontation data in the target game match.

[0130] In some embodiments of this application, a computer-readable storage medium is provided, storing a computer program that is loaded by a processor, causing the processor to perform the following steps:

[0131] Acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character;

[0132] Based on the analysis results of the game confrontation data, the confrontation tag information of the human-computer controlled character corresponding to the game confrontation data is determined;

[0133] The human-machine control role is controlled to output interactive information associated with the adversarial tag information to the user control role.

[0134] In some embodiments of this application, when the processor executes the computer program, it further performs the following steps: extracting historical interaction information from historical sample games; processing the historical interaction information based on a pre-trained language classification model to obtain tag information corresponding to the historical interaction information; associating the historical interaction information with the tag information and storing it in a preset database; querying the preset database to obtain target interaction information associated with the adversarial tag information; and controlling the human-machine control role to output the target interaction information to the user control role.

[0135] In some embodiments of this application, when the processor executes the computer program, it further performs the following steps: acquiring first sample interaction information and second sample interaction information carrying sample labels; training an initial language classification model based on the second sample interaction information and the sample labels corresponding to the second sample interaction information; processing the first sample interaction information based on the initial language classification model and outputting predicted label information corresponding to the first sample interaction information; updating the initial language classification model according to the first sample interaction information and the predicted label information corresponding to the first sample interaction information, the second sample interaction information and the sample labels corresponding to the second sample interaction information, to obtain a trained language classification model.

[0136] In some embodiments of this application, when the processor executes the computer program, it further implements the following steps: based on the parsing results of the character data corresponding to the human-computer controlled character and the character data corresponding to the user-controlled character in the game confrontation data, it determines the game event between the human-computer controlled character and the user-controlled character; based on the preset association relationship between the event and the emotion tag, it sets the emotion tag associated with the game event as the confrontation tag information corresponding to the human-computer controlled character in the game confrontation data.

[0137] In some embodiments of this application, when the processor executes the computer program, it further implements the following steps: controlling the human-machine control role to output facial expression interaction information associated with the adversarial tag information to the user control role; and in response to a feedback operation instruction executed by the user control role under the control of the target user, controlling the human-machine control role to output text interaction information associated with the adversarial tag information to the user control role.

[0138] In some embodiments of this application, when the processor executes the computer program, it also performs the following steps: controlling the human-machine control role to output text interaction information associated with the adversarial tag information and the feedback operation instructions to the user control role.

[0139] In some embodiments of this application, when the processor executes the computer program, it also performs the following steps: outputting target text interaction information associated with the role type in the text interaction information to the user control role based on the role type corresponding to the human-machine control role.

[0140] In some embodiments of this application, when the processor executes the computer program, it further implements the following steps: obtaining the feedback interval between the feedback operation instruction and the facial expression interaction information; if the feedback interval is less than a preset interval threshold, controlling the human-computer control role to output text interaction information associated with the adversarial tag information to the user control role.

[0141] In some embodiments of this application, when the processor executes the computer program, it also performs the following steps: obtaining the historical game matches of the user corresponding to each user-controlled character in the target game match; if the interaction frequency of the user in the historical game match meets the preset interaction conditions, then obtaining the game confrontation data in the target game match.

[0142] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0143] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] The foregoing has provided a detailed description of a game interaction simulation method, apparatus, computer device, and storage medium provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A game interaction simulation method, characterized in that: Acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character; Based on the analysis results of the game's competitive data, the competitive tag information corresponding to the human-computer controlled character in the game's competitive data is determined; the competitive tag information includes tag information obtained through sentiment analysis of users; The human-machine control role is controlled to output interactive information associated with the adversarial tag information to the user control role; Before the step of controlling the human-machine control role to output interactive information associated with the adversarial tag information to the user control role, the method further includes: Extract historical interaction information from historical sample games; the historical interaction information includes real interaction information between players in historical sample games. The historical interaction information is processed based on the pre-trained language classification model to obtain the label information corresponding to the historical interaction information; The historical interaction information is associated with the tag information and stored in a preset database; The control of the human-machine interface to output interactive information associated with the adversarial tag information to the user interface includes: Query the preset database to obtain target interaction information associated with the adversarial tag information; The human-machine interface controller outputs the target interaction information to the user interface controller.

2. The game interaction simulation method according to claim 1, characterized in that, The language classification model was pre-trained through the following steps: Acquire first sample interaction information and second sample interaction information carrying sample tags; the first sample interaction information includes unprocessed interaction information without sample tags, and the second sample interaction information includes interaction information preprocessed based on manual identification and carrying sample tags. An initial language classification model is trained based on the second sample interaction information and the sample labels corresponding to the second sample interaction information. The interaction information of the first sample is processed based on the initial language classification model, and the predicted label information corresponding to the interaction information of the first sample is output. Based on the first sample interaction information and the predicted label information corresponding to the first sample interaction information, the second sample interaction information and the sample label corresponding to the second sample interaction information, the initial language classification model is updated to obtain the trained language classification model.

3. The game interaction simulation method according to claim 1, characterized in that, The step of determining the adversarial tag information of the human-computer controlled character corresponding to the game adversarial data based on the analysis results of the game adversarial data includes: Based on the analysis results of the character data corresponding to the human-computer controlled character and the character data corresponding to the user-controlled character in the game confrontation data, the game events between the human-computer controlled character and the user-controlled character are determined; Based on the preset association between events and emotion tags, the emotion tags associated with the game events are set as the adversarial tag information of the game adversarial data corresponding to the human-computer control character.

4. The game interaction simulation method according to claim 1, characterized in that, The interactive information includes facial expression interaction information and text interaction information; The control of the human-machine interface to output interactive information associated with the adversarial tag information to the user interface includes: The human-machine control character is controlled to output facial expression interaction information associated with the adversarial tag information to the user control character; In response to a feedback operation command executed by the user control role under the control of the target user, the human-machine control role is controlled to output text interaction information associated with the adversarial tag information to the user control role.

5. The game interaction simulation method according to claim 4, characterized in that, The control of the human-machine interface to output text interaction information associated with the adversarial tag information to the user interface includes: The human-machine control role outputs text interaction information associated with the adversarial tag information and the feedback operation instructions to the user control role.

6. The game interaction simulation method according to claim 4, characterized in that, The control of the human-machine interface to output text interaction information associated with the adversarial tag information to the user interface includes: Based on the role type corresponding to the human-computer control role, the target text interaction information associated with the role type in the text interaction information is output to the user control role.

7. The game interaction simulation method according to claim 4, characterized in that, Before the step of controlling the human-machine control role to output text interaction information associated with the adversarial tag information to the user control role, the method further includes: Obtain the feedback interval between the feedback operation command and the corresponding facial expression interaction information; If the feedback interval is less than a preset interval threshold, then the step of controlling the human-machine control role to output text interaction information associated with the adversarial tag information to the user control role is executed.

8. The game interaction simulation method according to any one of claims 1 to 7, characterized in that, Before the step of acquiring game data from the target game match, the method further includes: Retrieve the historical game matches of the user corresponding to each user-controlled character in the target game; If the frequency of interaction of the user in the historical game matches meets the preset interaction conditions, then the step of obtaining the game confrontation data in the target game match is executed.

9. A game interaction simulation device, characterized in that, include: The acquisition module is used to acquire game confrontation data in a target game match, wherein the target game match includes at least one user-controlled character and one AI-controlled character; The parsing module is used to determine the adversarial tag information of the human-computer controlled character corresponding to the game adversarial data based on the parsing results of the game adversarial data; the adversarial tag information includes tag information obtained through user sentiment analysis; The output module is used to control the human-machine control role to output interactive information associated with the adversarial tag information to the user control role; Before the step of controlling the human-machine control role to output interactive information associated with the adversarial tag information to the user control role, the method further includes: Extract historical interaction information from historical sample games; the historical interaction information includes real interaction information between players in historical sample games. The historical interaction information is processed based on the pre-trained language classification model to obtain the label information corresponding to the historical interaction information; The historical interaction information is associated with the tag information and stored in a preset database; The control of the human-machine interface to output interactive information associated with the adversarial tag information to the user interface includes: Query the preset database to obtain target interaction information associated with the adversarial tag information; The human-machine interface controller outputs the target interaction information to the user interface controller.

10. A computer device, characterized in that, The computer device includes: One or more processors; Memory; and One or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the game interaction simulation method of any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, It stores a computer program, which is loaded by a processor to execute the game interaction simulation method according to any one of claims 1 to 8.

Images