AI-based anti-cheating detection method and system

Abstract

The present invention discloses an AI-based anti-cheating detection method and system, which relates to the field of anti-cheating detection technology. The key points of its technical solution include the following steps: inputting the player behavior characteristics of the current game match into a cheating prediction model to predict a cheating prediction result; if the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the cheating prediction results in the historical game match, a first anti-cheating processing solution is generated; if the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the cheating prediction results in the historical game match, the player behavior characteristics of the current game match are compared with the player behavior characteristics of the historical game match to obtain the distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and the different behavior characteristics is processed and analyzed to obtain a first behavior distribution feature and a second behavior distribution feature; the effect is that cheating behavior can be identified more comprehensively and accurately.

Description

Technical Field

[0001] The present invention relates to the field of anti-cheating detection technology, and more specifically, to an AI-based anti-cheating detection method and system. Background Art

[0002] Against the backdrop of today's booming digital entertainment industry, online games have become an important part of people's daily entertainment. From multiplayer online tactical competitive games to immersive massively multiplayer online role-playing games, various online games have attracted a large number of players. However, with the prosperity of the gaming industry, the problem of game cheating has become increasingly prominent, becoming a key obstacle to the healthy development of the gaming ecosystem.

[0003] With the development of gaming technology, cheating behaviors in games have become diverse and constantly evolving. Some cheaters exploit game loopholes to modify game client data or server interaction data to illegally obtain resources and quickly increase levels. Players may use loopholes to obtain large amounts of game currency or rare items to disrupt the economic balance within the game, or achieve instant level increases through data tampering, disrupting the normal growth system of the game and seriously affecting the fair competition environment of the game and the gaming experience of other players.

[0004] Currently, the gaming industry primarily employs a variety of anti-cheating technologies to address cheating. Common AI-based anti-cheating detection methods use pre-set rules to determine whether player behavior violates regulations. For example, rules such as upper limits on aiming speed and lower limits on skill release intervals are set. If a player's behavior exceeds these rules, it is considered cheating. However, this approach has limitations. For one thing, the rules are difficult to cover all possible cheating behaviors and special circumstances within normal gameplay, making it prone to misjudgments. Furthermore, cheaters can exploit loopholes in the rules and employ more subtle cheating methods to circumvent detection, rendering rule-based detection methods gradually ineffective. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide an AI-based anti-cheating detection method and system.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The AI-based anti-cheating detection method includes the following steps:

[0008] Step S1: Inputting the player behavior characteristics of the current game into the cheating prediction model to obtain a cheating prediction result;

[0009] Step S2: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, a first anti-cheating solution is generated;

[0010] Step S3: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the player behavior characteristics of the current game match and the historical game match are compared to obtain a distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and different behavior characteristics is processed and analyzed to obtain a first behavior distribution characteristic and a second behavior distribution characteristic;

[0011] Step S4: comparing the anti-cheating solution sets corresponding to the cheating prediction results based on the first behavior distribution feature and the second behavior distribution feature to extract overlapping solution sets and different solution sets;

[0012] Step S5: Analyze the variation characteristics between the different partial solutions in the set of different partial solutions to obtain a screening result set of different partial solutions, and integrate the screening result set of different partial solutions with the set of overlapping partial solutions to obtain a second anti-cheating processing solution.

[0013] Preferably, step S1 specifically includes the following steps:

[0014] Obtaining the game mode information and time period information of the current game match, and collecting historical cheating records of all historical game matches under the same conditions based on the game mode information and time period information;

[0015] Obtain the distribution of player behavior characteristics in the current game match, and filter out historical game matches with the same distribution of player behavior characteristics as the current game match from all historical game matches based on the distribution of player behavior characteristics;

[0016] Statistically marking feature information of each player's behavior and adjacent behaviors in historical game matches into a first feature information set, and establishing a cheating prediction model based on the first feature information set;

[0017] Obtain feature information of each player's behavior and adjacent behaviors in the current game and mark them as a second feature information set;

[0018] Behavioral feature information of a target player in the current game match is obtained, and the second feature information set and the behavioral feature information of the target player are input into a cheating prediction model to obtain a cheating prediction result.

[0019] Preferably, step S2 specifically includes the following steps:

[0020] If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, then the anti-cheating solution for the historical game match that resulted in cheating prediction results is extracted;

[0021] The anti-cheating solution is used as the anti-cheating solution for the cheating prediction results in the current game and is marked as the first anti-cheating solution.

[0022] Preferably, step S3 specifically includes the following steps:

[0023] If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the distribution of similar and different behavior characteristics between the historical game match and the current game match will be counted;

[0024] If similar behavior features are adjacent and concentrated, and different behavior features are adjacent and concentrated, then the first behavior distribution feature is output;

[0025] If the similar behavior features and the different behavior features are intersected and adjacent to each other, the second behavior distribution feature is output.

[0026] Preferably, step S4 specifically includes the following steps:

[0027] Extract the locations of suspected cheating behaviors in historical game matches from the cheating prediction results and mark them as the first type of behavior location information set;

[0028] Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the first behavior distribution feature and marking them as a second type of behavior location information set;

[0029] Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the second behavior distribution feature and marking them as the third type of behavior location information set;

[0030] Extracting and processing the first type anti-cheating solution set, the second type anti-cheating solution set, and the third type anti-cheating solution set corresponding to the cheating prediction results based on the first type behavior location information set, the second type behavior location information set, and the third type behavior location information set;

[0031] The first, second and third anti-cheating solution sets are compared to extract the anti-cheating solutions for the overlapping parts and the anti-cheating solutions for the different parts. All the anti-cheating solutions for the overlapping parts are combined to form an overlapping solution set, and all the anti-cheating solutions for the different parts are combined to form a different solution set.

[0032] Preferably, step S5 specifically includes the following steps:

[0033] According to the difference partial solution set, the main influencing factors of the anti-cheating solution decision-making based on the cheating prediction results are statistically analyzed to obtain the difference main decision factor set;

[0034] The distribution and development change direction of each main decision factor in the difference main decision factor set are processed to obtain the difference plan decision influence set;

[0035] Get the most recent historical anti-cheating solutions for the current game platform;

[0036] The historical anti-cheating solutions and the historical difference partial solution sets of the corresponding behaviors are processed to obtain a second proportion variation data set;

[0037] According to the second proportion variation data set and the difference scheme decision influence set, various difference partial schemes are screened out from the difference partial scheme set to obtain a difference partial scheme screening result set;

[0038] The second anti-cheating processing solution is obtained by integrating the screening result set of the difference part solution with the overlapping part solution set.

[0039] Preferably, the distribution and development change direction of each main decision factor in the difference main decision factor set are processed to obtain the difference solution decision impact set, which specifically includes the following steps:

[0040] Calculate the variation range of the distribution proportion value of each main decision factor in the difference main decision factor set to obtain a first proportion variation data set;

[0041] Evaluate the development and change direction of each main decision factor in the difference part of the plan set to obtain the difference plan development and change information set;

[0042] The decision impact coefficient between the first proportion value variation data set and the difference scheme development change information set is evaluated to obtain the difference scheme decision impact set.

[0043] Preferably, the historical anti-cheating solutions and the historical difference partial solution sets of the corresponding behaviors are processed to obtain a second percentage variation data set, which specifically includes the following steps:

[0044] Extracting the historical difference partial solution set corresponding to the behavior of the difference partial solution set from the historical anti-cheating solutions;

[0045] A second proportion variation data set corresponding to each difference partial solution in the historical difference partial solution set is extracted.

[0046] AI-based anti-cheating detection system, including:

[0047] Prediction module: Inputs the player behavior characteristics of the current game into the cheating prediction model to obtain the cheating prediction results;

[0048] Generation module: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, a first anti-cheating solution is generated;

[0049] Processing module: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the player behavior characteristics of the current game match and the historical game match are compared to obtain the distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and different behavior characteristics is processed and analyzed to obtain a first behavior distribution characteristic and a second behavior distribution characteristic;

[0050] Comparison module: compares the anti-cheating solution sets corresponding to the cheating prediction results based on the first behavior distribution feature and the second behavior distribution feature to extract overlapping solution sets and different solution sets;

[0051] Integration module: Analyze the change characteristics between each difference partial solution set to obtain a difference partial solution screening result set, and integrate the difference partial solution screening result set with the overlapping partial solution set to obtain a second anti-cheating processing solution.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] This application predicts cheating outcomes by inputting the behavioral characteristics of players in the current game match into a cheating prediction model. This model is built based on a large amount of historical game match data and can learn the subtle and complex characteristic differences between normal player behavior and cheating player behavior, greatly improving the accuracy of cheating behavior detection and reducing misjudgments and missed detections. At the same time, it processes the distribution of similar and different behavioral characteristics between current and historical game matches, considers player behavior from multiple dimensions, and also takes into account factors such as game stage and resource acquisition, avoiding the one-sidedness of single-dimensional analysis and identifying cheating behavior more comprehensively and accurately.

[0054] When the current game player behavior characteristics are consistent with historical cheating behavior characteristics, the historical anti-cheating solution is directly extracted to generate the first anti-cheating solution. This can quickly take effective measures against known cheating patterns without having to re-explore the response plan. For situations that are inconsistent with historical cheating behavior characteristics, suitable solutions are screened from the difference part of the solution set and integrated with the overlapping part of the solution set to generate the second anti-cheating solution. Taking the newly emerging cheating behavior of players exploiting game loopholes to gain special advantages as an example, by analyzing its unique behavioral characteristic distribution, blocking measures for the loophole and monitoring strategies for player behavior are screened out, and a solution is customized to specifically deal with the player's suspected cheating behavior, ensuring that anti-cheating measures are accurately adapted to different types of cheating situations.

[0055] This technical solution can promptly detect changes in the game environment and emerging cheating trends by continuously analyzing historical game data and historical anti-cheating solutions. The system can adjust cheating prediction models and anti-cheating strategies based on newly accumulated historical data and solutions to adapt to game changes and always maintain the effectiveness of anti-cheating. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a schematic diagram of the steps of the AI-based anti-cheating detection method proposed by the present invention;

[0057] Figure 2 This is a module diagram of the AI-based anti-cheating detection system proposed in the present invention. DETAILED DESCRIPTION

[0058] Reference Figures 1 to 2 .

[0059] Example 1 further explains the AI-based anti-cheating detection method and system proposed in the present invention.

[0060] The AI-based anti-cheating detection method includes the following steps:

[0061] Step S1: Inputting the player behavior characteristics of the current game into the cheating prediction model to obtain a cheating prediction result;

[0062] Step S2: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, a first anti-cheating solution is generated;

[0063] Step S3: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the player behavior characteristics of the current game match and the historical game match are compared to obtain a distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and different behavior characteristics is processed and analyzed to obtain a first behavior distribution characteristic and a second behavior distribution characteristic;

[0064] Step S4: comparing the anti-cheating solution sets corresponding to the cheating prediction results based on the first behavior distribution feature and the second behavior distribution feature to extract overlapping solution sets and different solution sets;

[0065] Step S5: Analyze the variation characteristics between the different partial solutions in the set of different partial solutions to obtain a screening result set of different partial solutions, and integrate the screening result set of different partial solutions with the set of overlapping partial solutions to obtain a second anti-cheating processing solution.

[0066] Consider a multiplayer online competitive game with a 5v5 team-based match mode and a time zone of 8:00 PM to 10:00 PM. First, the current game mode (5v5 team-based match) and time zone (8:00 PM to 10:00 PM) are obtained. Then, historical cheating records for all games played under the same game mode and time zone are collected in a database. For example, if cheating is detected in 100 games played during this time zone and game mode, where some players gained an unfair advantage by using cheats such as auto-aiming and cross-talk, the player behavior characteristics of the current game are then fed into a cheating prediction model to generate a cheating prediction result.

[0067] If the behavioral characteristics of player A in the current game match are found to be consistent with those of players predicted to have cheated in historical game matches, for example, if a player historically used an auto-aim cheat and exhibited attack frequency and skill release characteristics similar to player A, then the anti-cheating solution for that cheating prediction in historical game matches will be extracted from the database. For example, the anti-cheating solution previously implemented for players using auto-aim cheats was to directly ban their accounts for 7 days and publicly announce their cheating behavior in the game.

[0068] The extracted anti-cheating solution is used as the anti-cheating solution for the cheating prediction results of player A in the current game, and is marked as the first anti-cheating solution. That is, player A's account is also banned for 7 days, and his cheating behavior is announced in the game.

[0069] Suppose the behavioral characteristics of player B in the current game are inconsistent with those of players who were predicted to cheat in historical game matches. Player B's behavioral characteristics are cautious movement and efficient resource acquisition in the early game, but suddenly some aggressive moves in the late game that are inconsistent with their early style. The distribution of similar and different behavioral characteristics between historical game matches and the current game is statistically analyzed. For example, it is found that in historical games, some players are efficient in resource acquisition in the early game, but aggressive moves in the late game are rare. In these historical games, the similar behavioral characteristics are early resource acquisition, and the different behavioral characteristics are aggressive moves in the late game.

[0070] After determining the distribution characteristics of player B's first and second behaviors, we extract the anti-cheating solutions for the overlapping and diverging parts. We then combine all the overlapping solutions into a set of overlapping solutions, and all the diverging solutions into a set of diverging solutions. We then integrate the resulting set of diverging solutions with the set of overlapping solutions to create a second anti-cheating solution. For example, we might initially issue an in-game warning to player B, restrict some of their game functions (such as limiting their purchase of powerful items), and continuously monitor their subsequent behavior.

[0071] Step S1 specifically includes the following steps:

[0072] Obtain the game mode and time period of the current game, and use this information to compile statistics on cheating history for all previous games under the same conditions.

[0073] Obtain the distribution of player behavior characteristics in the current game match, and filter out historical game matches with the same distribution of player behavior characteristics as the current game match from all historical game matches based on the distribution of player behavior characteristics;

[0074] The feature information of each player's behavior and adjacent behaviors in the historical game matches is statistically marked as a first feature information set, and a cheating prediction model is established based on the first feature information set;

[0075] Obtain feature information of each player's behavior and adjacent behaviors in the current game and mark them as a second feature information set;

[0076] Behavioral feature information of a target player in the current game match is obtained, and the second feature information set and the behavioral feature information of the target player are input into a cheating prediction model to obtain a cheating prediction result.

[0077] First, the current game mode (5v5 team battle) and time period (8:00 PM - 10:00 PM) are obtained. Then, historical cheating records for all games played under the same game mode and time period are collected in the database. For example, cheating was found in 100 games during that time period and game mode, with some players gaining an unfair advantage by using cheat programs, such as auto-aim and perspective-seeking.

[0078] Obtain the distribution of player behavior characteristics in the current game, such as movement speed, attack frequency, and skill release timing. Suppose, in the current game, Player A's attack frequency is abnormally higher than average, and their skill release timing is extremely precise. Based on the distribution of these player behavior characteristics, select historical game matches from all historical game matches that share the same distribution of behavior characteristics as the current game match. For example, select 10 historical game matches where players also have high attack frequencies and precise skill releases.

[0079] The feature information of each player's behavior and adjacent behaviors in the selected historical game matches is statistically labeled as a first feature information set. For example, information such as the player's movement direction before attacking and the skill connection after attacking is recorded. Based on this first feature information set, a cheating prediction model is established. This model can be a deep learning-based neural network model. By learning from a large amount of historical data, it can understand the characteristic differences between normal player behavior and cheating player behavior.

[0080] The feature information of each player's behavior and adjacent behaviors in the current game match is obtained and marked as a second feature information set. Next, the feature information of the target player (e.g., Player A) in the current game match is obtained. This second feature information set and the target player's feature information are input into the cheating prediction model. After analysis and calculation by the model, a cheating prediction result is obtained, indicating that Player A has a high suspicion of cheating.

[0081] Step S2 specifically includes the following steps:

[0082] If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, then the anti-cheating solution for the historical game match that resulted in cheating prediction results is extracted;

[0083] The anti-cheating solution is used as the anti-cheating solution for the cheating prediction results in the current game and is marked as the first anti-cheating solution.

[0084] This application takes a game as an example. The game process monitoring system continuously collects its behavioral data, such as skill release frequency, movement accuracy, resource acquisition efficiency, etc., to form the player behavior characteristics of the current game.

[0085] During the game operation process, a large amount of historical game match data has been accumulated, and detailed records have been made of the games in which cheating was determined to have occurred, including the behavioral characteristics of the cheating players and the anti-cheating solutions adopted for the cheating behavior.

[0086] After the current game ends, the system compares the behavioral characteristics of the current game with the behavioral characteristics of players in the historical database who were predicted to have cheated. Suppose, in the historical database, Player C was previously identified as using an auto-aim cheat. Their behavioral characteristics manifested as an extremely high skill hit rate, accurately targeting key enemy heroes in complex teamfights, and skillfully avoiding enemy skills with their positioning. Player D, in this game, also demonstrated an extremely high skill hit rate, with nearly every skill cast against key enemy heroes in teamfights. Their positioning was also incredibly precise, and they were almost never hit by enemy skills. The system determined that Player D's behavioral characteristics in the current game match matched those of Player C in the historical game that was predicted to have cheated.

[0087] At this point, the system extracts the anti-cheating solution for Player C's cheating behavior from the historical records. At the time, the game authorities had banned Player D for 14 days and publicly announced their cheating behavior. The system extracts this anti-cheating solution and uses it as the anti-cheating solution for Player C's cheating prediction results in the current game, marking it as the first anti-cheating solution.

[0088] Step S3 specifically includes the following steps:

[0089] If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the distribution of similar and different behavior characteristics between the historical game match and the current game match will be counted;

[0090] If similar behavior features are adjacent and concentrated, and different behavior features are adjacent and concentrated, then the first behavior distribution feature is output;

[0091] If the similar behavior features and the different behavior features are intersected and adjacent to each other, the second behavior distribution feature is output.

[0092] Player B participated in a game competition. During the game, player B's behavioral data was continuously collected, including shooting frequency, aiming speed, movement trajectory, material pickup preference and other behavioral characteristics.

[0093] The database stores a large amount of historical game data, including records of players who were found to have cheated in some of these games. The system compares Player B's behavior in the current game with the behavior of players who were predicted to have cheated in these historical games and finds that Player B's behavior is inconsistent with these historical cheating behavior characteristics.

[0094] Next, the system begins to analyze the distribution of similar and different behavioral characteristics between historical game matches and the current game (the one in which Player B participates). For example, in historical game matches, normal players typically move cautiously in the early stages of the game, prioritizing the search for basic supplies to ensure their survival. Player B also moves cautiously in the early stages of the game, focusing on collecting basic supplies such as weapons and ammunition. This is a similar behavioral characteristic.

[0095] In historical matches, a typical player, upon entering the finals, would choose the appropriate time to move and ambush based on the location of the safe zone and the distribution of remaining enemies. However, after entering the finals, Player B's movement patterns and decision-making are quite unique. He deviates from common strategies and always takes risks, choosing to attack where the enemy might be hiding. This unique behavior is a differential behavioral characteristic.

[0096] First Behavior Distribution Feature: If similar behavioral traits of player B (e.g., cautiously collecting resources in the early stages) cluster within a certain period of time during the early stages of the game, adjacent behaviors are all centered around collecting resources. Distinctive behavioral traits (e.g., unique movement decisions in the final round) also cluster within this specific phase, and adjacent behaviors are all related and unique decisions. In other words, if similar behavioral traits are adjacent and clustered, and dissimilar behavioral traits are adjacent and clustered, the system outputs the first behavior distribution feature.

[0097] Second Behavior Distribution: If player B's gameplay shows similar behavior patterns (such as collecting resources) and different behavior patterns (such as unique movement decisions) intersect and overlap, for example, while collecting resources, they might suddenly make a few unique, adventurous moves to test the enemy's position, then resume collecting resources, and repeat this process. In this case, the system outputs a second behavior distribution pattern.

[0098] Step S4 specifically includes the following steps:

[0099] Extract the locations of suspected cheating behaviors in historical game matches from the cheating prediction results and mark them as the first type of behavior location information set;

[0100] Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the first behavior distribution feature and marking them as a second type of behavior location information set;

[0101] Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the second behavior distribution feature and marking them as the third type of behavior location information set;

[0102] Extracting and processing the first type anti-cheating solution set, the second type anti-cheating solution set, and the third type anti-cheating solution set corresponding to the cheating prediction results based on the first type behavior location information set, the second type behavior location information set, and the third type behavior location information set;

[0103] The first, second and third anti-cheating solution sets are compared to extract the anti-cheating solutions for the overlapping parts and the anti-cheating solutions for the different parts. All the anti-cheating solutions for the overlapping parts are combined to form an overlapping solution set, and all the anti-cheating solutions for the different parts are combined to form a different solution set.

[0104] The system searches historical game data for locations similar to Player B's suspected cheating behavior. For example, Player B suddenly displays an extremely high accuracy rate with wall-penetrating shots at the 20-minute mark, suggesting the player is using a perspective cheat. The system searches historical cheating game data for similar wall-penetrating shooting behaviors and finds that this type of cheating often occurs between the 15-25 minute mark and is concentrated in specific map areas, such as the "Abandoned Factory" and "Old Town." These behavior locations are marked as the first category of behavior location information.

[0105] Based on player B's first behavior distribution feature, the system extracts the locations of his suspected cheating behaviors within that feature. Assume that player B's suspected cheating behaviors in the first behavior distribution feature are primarily concentrated in the mid-game resource competition phase, with frequent and unusually accurate shooting near specific resource points. The system then labels this behavior location information, such as the specific resource point name and the time range of that phase, as the second category of behavior location information.

[0106] Based on player B's second behavior distribution, the system identifies the locations of suspected cheating behaviors. For example, in player B's second behavior distribution, their suspected cheating behaviors appear interspersed with normal resource gathering and transfers, with unusual aiming and shooting behaviors occurring randomly in different map areas. The system records the map coordinates and approximate time of these behaviors, marking them as the third type of behavior location information set.

[0107] Based on the first set of behavior location information—the location information of similar cheating behaviors in historical game matches—the system extracts historical anti-cheating solutions for these cheating behaviors and locations. For example, historically, solutions for wall-penetrating shooting cheats that occurred during specific time periods in "Abandoned Factories" and "Old Town" included real-time monitoring of player behavior in those areas and temporary restrictions on suspected cheaters (such as reduced movement speed and weapon usage restrictions). These solutions constitute the first set of anti-cheating solutions.

[0108] Based on the second set of behavior location information, we extract anti-cheating solutions targeting the locations of suspected cheating behaviors of player B in the first set of behavior distribution features. For example, for unusually accurate shooting behavior near a specific resource point in the middle of the game, anti-cheating solutions include increasing field of view monitoring in that area and issuing warnings to players who frequently engage in unusual behavior in that area. These solutions constitute the second set of anti-cheating solutions.

[0109] Based on the third category of behavior location information, we extract anti-cheating solutions corresponding to the locations of player B's suspected cheating behaviors within the second category of behavior distribution. For example, for abnormal aiming and shooting behaviors that randomly occur during the collection and transfer of supplies, we can adopt solutions such as dynamically adjusting monitoring parameters and tracking and analyzing the suspicious player's behavior trajectory. These solutions form the third category of anti-cheating solutions.

[0110] We found that some solutions appeared in all three sets of solutions, such as the behavioral data record analysis of suspected cheating players. These overlapping anti-cheating solutions were combined to form an overlapping solution set.

[0111] At the same time, there are also solutions that appear in one or two categories of solutions (such as the temporary restriction of operations on specific areas that is unique to the first category of solutions). These different anti-cheating solutions are combined into a different solution set.

[0112] Step S5 specifically includes the following steps:

[0113] According to the difference partial solution set, the main influencing factors of the anti-cheating solution decision-making based on the cheating prediction results are statistically analyzed to obtain the difference main decision factor set;

[0114] The distribution and development change direction of each main decision factor in the difference main decision factor set are processed to obtain the difference plan decision influence set;

[0115] Get the most recent historical anti-cheating solutions for the current game platform;

[0116] The historical anti-cheating solutions and the historical difference partial solution sets of the corresponding behaviors are processed to obtain a second proportion variation data set;

[0117] According to the second proportion variation data set and the difference scheme decision influence set, various difference partial schemes are screened out from the difference partial scheme set to obtain a difference partial scheme screening result set;

[0118] The second anti-cheating processing solution is obtained by integrating the screening result set of the difference part solution with the overlapping part solution set.

[0119] Based on the difference partial solution set, the main influencing factors of anti-cheating solution decision-making based on the cheating prediction results are counted. For example, in the difference partial solution set of player B, the main influencing factors include the economic status in the game (such as having a large amount of game currency to purchase special props), the level of the game character (a high level may unlock special abilities), and the game time (cheating behavior is more frequent at certain specific time points). These main influencing factors are sorted out to obtain the difference main decision factor set.

[0120] The first percentage variation dataset is calculated by calculating the variation in the distribution percentage of each main decision factor in the variance main decision factor set. For example, it was found that the ratio of game currency acquired by player B to total acquired currency in the early stages of the game varied significantly, indicating a significant difference from the data of normal players.

[0121] The development and change direction of each major decision-making factor in the difference partial solution set is evaluated to obtain the difference solution development and change information set. For example, it is observed that the level of player B's game character suddenly increases in the middle of the game, and the frequency and type of props purchased with game currency also change abnormally.

[0122] The decision-making influence coefficients of the first percentage value variation dataset and the difference solution development information set are evaluated to obtain the difference solution decision-making influence set. By combining this data, we can determine which factors have the greatest impact on player B's suspected cheating behavior and how these factors interact to influence the anti-cheating solution decision.

[0123] Obtain the current gaming platform's recent historical anti-cheating measures. For example, the gaming platform has taken measures such as temporarily banning accounts and restricting in-game transactions in response to similar suspected cheating behaviors in the past month.

[0124] Extract historical difference partial solution sets corresponding to the behaviors in the difference partial solution set from historical anti-cheating solutions. For example, find anti-cheating solutions that have historically addressed similar players' abnormal behavior in the game economy and character level. Then, extract the corresponding second percentage change data set for each difference partial solution in the historical difference partial solution set. For example, analyze the historical changes in relevant data (such as the percentage of players' game currency holdings, the percentage of character level increase speeds, etc.) before and after the implementation of different anti-cheating measures in similar situations.

[0125] Based on the second percentage variation dataset and the difference solution decision impact set, various difference solutions are screened from the difference solution set. For example, if it is found that historically, measures restricting in-game transactions have been effective in addressing similar player economic anomalies, and the difference solution decision impact set also indicates that game economic factors have a significant impact on player B's suspected cheating behavior, then the option of restricting in-game transactions is screened out, resulting in a difference solution screening result set.

[0126] Integrate the solution sets from the different solutions with the overlapping solutions. Assume that the overlapping solution set includes real-time monitoring of player behavior, while the different solution set includes restricting in-game transactions for player B. Combining these two measures creates a second anti-cheating solution for player B: real-time behavior monitoring and restrictions on in-game transactions.

[0127] The distribution and development change direction of each main decision factor in the difference main decision factor set are processed to obtain the difference scheme decision impact set, which specifically includes the following steps:

[0128] Calculate the variation range of the distribution proportion value of each main decision factor in the difference main decision factor set to obtain a first proportion variation data set;

[0129] Evaluate the development and change direction of each main decision factor in the difference part of the plan set to obtain the difference plan development and change information set;

[0130] The decision impact coefficient between the first proportion value variation data set and the difference scheme development change information set is evaluated to obtain the difference scheme decision impact set.

[0131] In this game, a normal player spends the first 20 minutes gathering resources and cautiously developing, spending approximately 30%-40% of their total game time in combat. However, Player B actively engages in combat within the first 20 minutes, battling far more frequently than a normal player, with 60% of their game time spent in combat. Later in the game, while a normal player's combat time increases to 60%-70%, Player B's combat time drops to 40%. Calculations show that the percentage of Player B's game time spent in combat fluctuates by 20%-30% relative to the normal range in the early stages, and by -20%--30% in the later stages. These data constitute part of the first percentage variation dataset for the factor of game time.

[0132] Typically, a normal player's resource acquisition in the early and middle stages of the game accounts for 50%-60% of their total acquisition. Player B's resource acquisition in the early and middle stages accounts for as much as 80% of their total acquisition, but this percentage drops significantly to 20% in the later stages. Based on this calculation, the variation in Player B's resource acquisition percentage relative to the normal range in the early and middle stages is 20%-30%, and in the later stages it varies by -30%--40%, becoming the relevant data for the resource acquisition factor in the first percentage variation dataset.

[0133] A normal player's kill efficiency is low in the early game, with kills accounting for approximately 10%-20% of their total kills. Player B's kills account for 40% of their total kills in the early game, 40% in the mid-game, and 20% in the late game. Calculations show that Player B's kill efficiency percentage fluctuates by 20%-30% relative to the normal range in the early game, remains relatively stable in the mid-game, and fluctuates by -10%--20% in the late game. By summarizing these variations in kill efficiency, we obtain a complete dataset of the first percentage variation.

[0134] Player B's time usage shifts from aggressive early-game combat to less aggressive later in the game. This change suggests that their strategy may not be the typical gradual progression of a player. Instead, they exploit unknown advantages to quickly establish an advantage early on, then adjust their strategy later. The system records this shift in time allocation from aggressive to conservative, and its impact on their gaming behavior at different stages, such as the accumulation of advantageous resources through frequent combat in the early stages. This information focuses on the time factor in the development of different strategies.

[0135] Player B's resource acquisition is significantly higher than normal in the early stages, but then plummets in the later stages. This change may indicate that they acquired resources through illicit means, but later their access to resources became blocked or they no longer needed large amounts of resources. This initial high-to-low resource acquisition change can influence Player B's subsequent game behavior, such as leveraging the large amount of resources acquired early on to gain a combat advantage in the mid-game. This information provides information on the resource acquisition factor in the information set for the development and change of different strategies.

[0136] Player B's kill efficiency was unusually high in the early stages, maintained at a certain level in the mid-game, and then decreased in the later stages. This may indicate that they used some kind of cheating method to increase their kill rate in the early stages, adjusted their strategy in the mid-game due to changes in the game situation, but still maintained a certain advantage, which gradually weakened in the later stages. The system evaluates these periodic changes in kill efficiency, the possible reasons behind them, and their impact on the game process, forming the kill efficiency factor in the differential strategy development and change information set. The evaluation results of various factors are combined to form a complete differential strategy development and change information set.

[0137] A correlation analysis was performed between the first percentage variation dataset and the information set on the development and change of the different scenarios. It was found that player B's high early game time spent in combat was strongly correlated with high resource acquisition and kill efficiency, suggesting that he may be using cheating methods to quickly acquire resources and exploit his advantages in combat. Evaluation revealed that the influence coefficient of game time on his decision-making regarding suspected cheating behavior was 0.4, the influence coefficient of resource acquisition was 0.3, and the influence coefficient of kill efficiency was 0.3. By integrating these factors and their influence coefficients, we obtained the influence set for the different scenarios' decision-making.

[0138] The historical anti-cheating solutions and the historical difference partial solution sets of the corresponding behaviors are processed to obtain a second percentage variation dataset, specifically including the following steps:

[0139] Extracting the historical difference partial solution set corresponding to the behavior of the difference partial solution set from the historical anti-cheating solutions;

[0140] A second proportion variation data set corresponding to each difference partial solution in the historical difference partial solution set is extracted.

[0141] The game operator maintains a large collection of historical anti-cheating solutions. First, the system extracts from these historical anti-cheating solutions the historical divergent solutions corresponding to the behaviors in Player B's divergent solution set. For example, suppose Player B's divergent solution set includes unusually rapid character leveling, unusually large amounts of in-game currency acquisition, and frequent completion of difficult tasks in a short period of time. The system then searches the historical anti-cheating solutions for records related to these behaviors. For example, if it discovers that other players have similarly experienced unusually rapid character leveling in a certain period of time, the anti-cheating measures implemented at that time might include limiting the player's experience gain rate and closely monitoring their leveling behavior. For unusually large amounts of in-game currency acquisition, measures such as freezing in-game currency assets and investigating transaction records were implemented. For frequently completing difficult tasks in a short period of time, measures such as suspending task system access and verifying task completion paths were implemented. These historical anti-cheating solutions for similar behaviors are then compiled to form a historical divergent solution set corresponding to Player B's divergent solution set.

[0142] After obtaining the set of historically different partial scenarios, we further extracted the corresponding second percentage variation dataset for each of these scenarios. For the scenario involving limiting the experience gain rate, we calculated the change in the proportion of the player's character level increase to their total level increase before and after the implementation of this scenario. Assume that before the restriction, the player's character level increase in a week accounted for 80% of their total level increase. After the restriction, this proportion dropped to 20%. The change in the level increase percentage was calculated to be -60%. This change data was recorded for different time periods (e.g., daily and weekly) to form part of the second percentage variation dataset for the experience gain rate restriction scenario.

[0143] For the plan to freeze in-game currency assets and investigate transaction records, analyze the change in the proportion of player in-game currency holdings to the total in-game currency before and after implementation. For example, if player in-game currency holdings were 10% before implementation, after the freeze and investigation, the proportion dropped to 1%, resulting in a -9% change in the proportion of in-game currency holdings. Similarly, record the change at different time points to construct a second proportion change dataset corresponding to this plan.

[0144] For the solution that suspended access to the task system and verified task completion paths, we calculated the change in the percentage of high-difficulty tasks completed by players relative to the total number of completed tasks before and after implementation. If the percentage of high-difficulty tasks completed by a player was 50% before the implementation and dropped to 10% afterward, the change in percentage would be -40%. Similarly, we extracted and organized the percentage change data for each of the historically different solutions before and after implementation, ultimately creating a complete second percentage change dataset.

[0145] Example 2, an AI-based anti-cheating detection system, including:

[0146] Prediction module: Inputs the player behavior characteristics of the current game into the cheating prediction model to obtain the cheating prediction results;

[0147] Generation module: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, a first anti-cheating solution is generated;

[0148] Processing module: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the player behavior characteristics of the current game match and the historical game match are compared to obtain the distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and different behavior characteristics is processed and analyzed to obtain a first behavior distribution characteristic and a second behavior distribution characteristic;

[0149] Comparison module: compares the anti-cheating solution sets corresponding to the cheating prediction results based on the first behavior distribution feature and the second behavior distribution feature to extract overlapping solution sets and different solution sets;

[0150] Integration module: Analyze the change characteristics between each difference partial solution set to obtain a difference partial solution screening result set, and integrate the difference partial solution screening result set with the overlapping partial solution set to obtain a second anti-cheating processing solution.

[0151] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0152] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. The AI-based anti-cheating detection method is characterized by: The method comprises the following steps: Step S1: Inputting the player behavior characteristics of the current game into the cheating prediction model to obtain a cheating prediction result; Step S2: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, a first anti-cheating solution is generated; Step S3: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the player behavior characteristics of the current game match and the historical game match are compared to obtain a distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and different behavior characteristics is processed and analyzed to obtain a first behavior distribution characteristic and a second behavior distribution characteristic; Step S4: comparing the anti-cheating solution sets corresponding to the cheating prediction results based on the first behavior distribution feature and the second behavior distribution feature to extract overlapping solution sets and different solution sets; Extract the locations of suspected cheating behaviors in historical game matches from the cheating prediction results and mark them as the first type of behavior location information set; Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the first behavior distribution feature and marking them as a second type of behavior location information set; Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the second behavior distribution feature and marking them as the third type of behavior location information set; Extracting and processing the first type anti-cheating solution set, the second type anti-cheating solution set, and the third type anti-cheating solution set corresponding to the cheating prediction results based on the first type behavior location information set, the second type behavior location information set, and the third type behavior location information set; Compare the first, second, and third anti-cheating solution sets to extract overlapping anti-cheating solutions and different anti-cheating solutions. Combine all overlapping anti-cheating solutions to form an overlapping solution set, and combine all different anti-cheating solutions to form a different solution set. Step S5: Analyze the variation characteristics between the different partial solutions in the set of different partial solutions to obtain a screening result set of different partial solutions, and integrate the screening result set of different partial solutions with the set of overlapping partial solutions to obtain a second anti-cheating processing solution.

2. The AI-based anti-cheating detection method according to claim 1, characterized in that: Step S1 specifically includes the following steps: Statistically marking feature information of each player's behavior and adjacent behaviors in historical game matches into a first feature information set, and establishing a cheating prediction model based on the first feature information set; Obtain feature information of each player's behavior and adjacent behaviors in the current game and mark them as a second feature information set; Behavioral feature information of a target player in the current game match is obtained, and the second feature information set and the behavioral feature information of the target player are input into a cheating prediction model to obtain a cheating prediction result.

3. The AI-based anti-cheating detection method according to claim 1, characterized in that: Step S2 specifically includes the following steps: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, then the anti-cheating solution for the historical game match that resulted in cheating prediction results is extracted; The anti-cheating solution is used as the anti-cheating solution for the cheating prediction results in the current game and is marked as the first anti-cheating solution.

4. The AI-based anti-cheating detection method according to claim 1, characterized in that: Step S3 specifically includes the following steps: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the distribution of similar and different behavior characteristics between the historical game match and the current game match will be counted; If similar behavior features are adjacent and concentrated, and different behavior features are adjacent and concentrated, then the first behavior distribution feature is output; If the similar behavior features and the different behavior features are intersected and adjacent to each other, the second behavior distribution feature is output.

5. The AI-based anti-cheating detection method according to claim 4, characterized in that: Step S5 specifically includes the following steps: According to the difference partial solution set, the main influencing factors of the anti-cheating solution decision-making based on the cheating prediction results are statistically analyzed to obtain the difference main decision factor set; The distribution and development change direction of each main decision factor in the difference main decision factor set are processed to obtain the difference plan decision influence set; Get the most recent historical anti-cheating solutions for the current game platform; The historical anti-cheating solutions and the historical difference partial solution sets of the corresponding behaviors are processed to obtain a second proportion variation data set; According to the second proportion variation data set and the difference scheme decision influence set, various difference partial schemes are screened out from the difference partial scheme set to obtain a difference partial scheme screening result set; The second anti-cheating processing solution is obtained by integrating the screening result set of the difference part solution with the overlapping part solution set.

6. The AI-based anti-cheating detection method according to claim 5, characterized in that: The distribution and development change direction of each main decision factor in the difference main decision factor set are processed to obtain the difference scheme decision impact set, which specifically includes the following steps: Calculate the variation range of the distribution proportion value of each main decision factor in the difference main decision factor set to obtain a first proportion variation data set; Evaluate the development and change direction of each main decision factor in the difference part of the plan set to obtain the difference plan development and change information set; The decision impact coefficient between the first proportion value variation data set and the difference scheme development change information set is evaluated to obtain the difference scheme decision impact set.

7. The AI-based anti-cheating detection method according to claim 6, characterized in that: The historical anti-cheating solutions and the historical difference partial solution sets of the corresponding behaviors are processed to obtain a second percentage variation dataset, specifically including the following steps: Extracting the historical difference partial solution set corresponding to the behavior of the difference partial solution set from the historical anti-cheating solutions; A second proportion variation data set corresponding to each difference partial solution in the historical difference partial solution set is extracted.

8. An AI-based anti-cheating detection system, applied to the AI-based anti-cheating detection method according to any one of claims 1 to 7, characterized in that: include: Prediction module: Inputs the player behavior characteristics of the current game into the cheating prediction model to obtain the cheating prediction results; Generation module: If the player behavior characteristics of the current game match are consistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, a first anti-cheating solution is generated; Processing module: If the player behavior characteristics of the current game match are inconsistent with the player behavior characteristics of the historical game match that resulted in cheating prediction results, the player behavior characteristics of the current game match and the historical game match are compared to obtain the distribution of similar behavior characteristics and different behavior characteristics, and the distribution of the similar behavior characteristics and different behavior characteristics is processed and analyzed to obtain a first behavior distribution characteristic and a second behavior distribution characteristic; Comparison module: compares the anti-cheating solution sets corresponding to the cheating prediction results based on the first behavior distribution feature and the second behavior distribution feature to extract overlapping solution sets and different solution sets; Extract the locations of suspected cheating behaviors in historical game matches from the cheating prediction results and mark them as the first type of behavior location information set; Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the first behavior distribution feature and marking them as a second type of behavior location information set; Extracting the behavior locations of the suspected cheating behaviors in the cheating prediction results in the second behavior distribution feature and marking them as the third type of behavior location information set; Extracting and processing the first type anti-cheating solution set, the second type anti-cheating solution set, and the third type anti-cheating solution set corresponding to the cheating prediction results based on the first type behavior location information set, the second type behavior location information set, and the third type behavior location information set; Compare the first, second, and third anti-cheating solution sets to extract overlapping anti-cheating solutions and different anti-cheating solutions. Combine all overlapping anti-cheating solutions to form an overlapping solution set, and combine all different anti-cheating solutions to form a different solution set. Integration module: Analyze the change characteristics between each difference partial solution set to obtain a difference partial solution screening result set, and integrate the difference partial solution screening result set with the overlapping partial solution set to obtain a second anti-cheating processing solution.

Images