Anomaly recognition method and system based on artificial intelligence
By obtaining and analyzing the path data of intrusion detection events, identifying target intrusion knowledge points that are different from the basic intrusion knowledge points, and using prior labels to determine abnormal labels, it solves the accuracy and efficiency problems of traditional methods when detecting complex network attacks. Improved safety protection capabilities.
Patent Information
- Application Number
- CN202411166106.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
When facing complex and changing cyber attacks, traditional intrusion detection methods have low detection accuracy and high false alarm rates, making it difficult to identify mutated, new or unknown intrusion behaviors, and rely on manual analysis to be inefficient.
By obtaining the intrusion path data of the intrusion detection event, the basic intrusion knowledge points are determined, and the target intrusion knowledge points that are different from the basic intrusion knowledge points in the second intrusion data area are identified, and a priori intrusion derived exception tags are used. Determine the abnormal label derived from event intrusion to improve the accuracy of label determination.
It realizes accurate identification and label determination of complex network attacks, improves the overall security protection capability of the security protection platform, and reduces the dependence of manual analysis.
Smart Images

Figure CN119094181A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to an abnormality recognition method and system based on artificial intelligence. Background Art
[0002] With the rapid development of information technology, network security issues have become increasingly prominent, and various network attack methods have emerged in an endless stream, seriously threatening the information security of enterprises and individuals. In order to cope with these threats, security protection platforms have emerged, which use technical means such as intrusion detection systems (IDS) to monitor network traffic in real time and promptly detect and deal with potential intrusions.
[0003] However, traditional intrusion detection methods often have problems with low detection accuracy and high false alarm rate when facing complex and changeable network attacks. This is mainly because traditional methods usually match based on known attack patterns and are difficult to effectively identify variant, new or unknown intrusion behaviors. In addition, traditional methods often rely on manual analysis and experience to determine the derived abnormal labels of intrusion events, which is inefficient and prone to errors. Summary of the invention
[0004] In view of this, the purpose of this application is to provide an abnormality identification method and system based on artificial intelligence.
[0005] In conjunction with the first aspect of the present application, there is provided an artificial intelligence-based anomaly recognition method, which is applied to an artificial intelligence-based anomaly recognition system, and the method comprises: Acquire first intrusion path data of an intrusion detection event in the security protection platform, wherein the first intrusion path data represents an intrusion execution operation sequence of the corresponding intrusion detection event in the first intrusion data area; Determine a plurality of basic intrusion knowledge points based on the first intrusion path data; Acquire second intrusion path data of a target intrusion detection event in the security protection platform, wherein the second intrusion path data represents an intrusion execution operation sequence of the corresponding target intrusion detection event in a second intrusion data area, wherein the target intrusion detection event is an intrusion detection event in which a variation occurs in the intrusion execution operation sequence in the second intrusion data area, and the second intrusion data area is an intrusion data area after a service node of the first intrusion data area; Based on the second intrusion path data and the basic intrusion knowledge point, determining a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area; Based on the prior intrusion-derived anomaly label corresponding to the target intrusion knowledge point, an event intrusion-derived anomaly label of the target intrusion detection event is determined.
[0006] In a possible implementation of the first aspect, determining, based on the second intrusion path data and the basic intrusion knowledge point, a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area includes: Determine, based on the second intrusion path data, second intrusion description vector sequences corresponding to the target intrusion detection events respectively; Determine the deviation between the second intrusion description vector sequence and the intrusion description vector sequence corresponding to the basic intrusion knowledge point, and use the second intrusion description vector sequence with a deviation greater than a preset deviation as a candidate intrusion description vector sequence; generating at least one second intrusion description group according to grouping the candidate intrusion description vector sequence; The target intrusion knowledge point is determined based on a candidate intrusion description vector sequence covered by the second intrusion description group.
[0007] In a possible implementation of the first aspect, determining the event intrusion derived anomaly label of the target intrusion detection event based on the prior intrusion derived anomaly label corresponding to the target intrusion knowledge point includes: Determining a priori intrusion-derived anomaly label of the target intrusion knowledge point based on second intrusion path data corresponding to the second intrusion description group; If a key intelligence intrusion knowledge point whose prior intrusion derived anomaly label is an intelligence trigger label is determined from the target intrusion knowledge point, the target intrusion detection event associated with the second intrusion description group corresponding to the key intelligence intrusion knowledge point is determined as a key intelligence intrusion detection event.
[0008] In a possible implementation manner of the first aspect, the step of grouping the candidate intrusion description vector sequences to generate at least one second intrusion description group includes: Performing balanced extraction on the candidate intrusion description vector sequences to generate intrusion description vector extraction sequences, wherein the number of the intrusion description vector extraction sequences is less than the number of the candidate intrusion description vector sequences; The intrusion description vector extraction sequence is grouped to generate at least one second intrusion description group.
[0009] In a possible implementation of the first aspect, the candidate vector sequence is one of the grouped intrusion description vector sequences corresponding to the basic intrusion knowledge points, and the step of determining the preset deviation includes: determining a candidate deviation between the second intrusion description vector sequence and the candidate vector sequence; Based on a descending set of threshold values, sequentially determining the number of sequences of the second intrusion description vector sequences whose candidate deviations meet the candidate threshold values in the descending set; If the difference in the number of sequences between the number of sequences at the xth candidate threshold value and the number of sequences at the x+1th candidate threshold value meets the set requirement, the preset deviation is determined based on the xth candidate threshold value.
[0010] In a possible implementation of the first aspect, determining a plurality of basic intrusion knowledge points based on the first intrusion path data includes: Determine, based on the first intrusion path data, first intrusion description vector sequences corresponding to the intrusion detection events in the security protection platform; generating a plurality of first intrusion description groups according to grouping the first intrusion description vector sequence; Based on the first intrusion description vector sequence covered by the first intrusion description group, a plurality of basic intrusion knowledge points respectively corresponding to the plurality of first intrusion description groups are determined.
[0011] In a possible implementation of the first aspect, determining, based on the first intrusion description vector sequence covered by the first intrusion description group, a plurality of basic intrusion knowledge points respectively corresponding to the plurality of first intrusion description groups includes: Extracting from a first intrusion description vector sequence covered by the first intrusion description group, and determining a group intrusion description vector sequence based on the extracted first intrusion description vector sequence; The group intrusion description vector sequences corresponding to the plurality of first intrusion description groups are determined as the intrusion description vector sequences of the plurality of basic intrusion knowledge points.
[0012] In a possible implementation manner of the first aspect, a regional service span corresponding to the second intrusion data area is equal to a regional service span corresponding to the first intrusion data area.
[0013] In a possible implementation of the first aspect, the candidate intrusion knowledge point is the basic intrusion knowledge point or the target intrusion knowledge point, and in the process of determining, based on the second intrusion path data and the basic intrusion knowledge point, a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area, the method further includes: If the number of pairings between the candidate intrusion knowledge point and the second intrusion path data is less than a set number, the candidate intrusion knowledge point is removed.
[0014] In combination with the second aspect of the present application, an artificial intelligence-based anomaly recognition system is provided, wherein the artificial intelligence-based anomaly recognition system includes a machine-readable storage medium and a processor, wherein the machine-readable storage medium stores machine-executable instructions, and when the processor executes the machine-executable instructions, the artificial intelligence-based anomaly recognition system implements the aforementioned artificial intelligence-based anomaly recognition method.
[0015] In conjunction with the third aspect of the present application, a computer-readable storage medium is provided, in which computer-executable instructions are stored. When the computer-executable instructions are executed, the aforementioned artificial intelligence-based anomaly identification method is implemented.
[0016] In combination with any of the above aspects, by obtaining the first intrusion path data of the intrusion detection event in the security protection platform, and determining multiple basic intrusion knowledge points based on these first intrusion path data, by obtaining the second intrusion path data of the target intrusion detection event, and combining the basic intrusion knowledge points, it is possible to accurately identify the target intrusion knowledge points that are different from the basic intrusion knowledge points in the second intrusion data area, and determine the event intrusion derived anomaly labels of the target intrusion detection event based on the prior intrusion derived anomaly labels corresponding to the target intrusion knowledge points, which not only depends on the current intrusion behavior characteristics, but also combines historical information and prior knowledge, thereby improving the accuracy of label determination. Through accurate identification of intrusion behavior and label determination, it can help the security protection platform to more effectively respond to various network security threats and improve the overall security protection capabilities of the system. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained by combining these drawings without paying creative work.
[0018] Figure 1 A flowchart of an anomaly recognition method based on artificial intelligence provided in an embodiment of the present application. DETAILED DESCRIPTION
[0019] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. According to the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0020] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or ends.
[0021] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0022] Figure 1 The flowchart of the abnormality identification method based on artificial intelligence provided by the embodiment of the present application is shown. It should be understood that in other embodiments, the order of some steps in the abnormality identification method based on artificial intelligence of this embodiment can be shared with each other according to actual needs, or some steps can be omitted or maintained. The abnormality identification method based on artificial intelligence includes: Step S110, obtaining first intrusion path data of an intrusion detection event in the security protection platform, wherein the first intrusion path data represents an intrusion execution operation sequence of the corresponding intrusion detection event in the first intrusion data area.
[0023] In this embodiment, the server extracts the first intrusion path data of the intrusion detection event from the security protection platform. These first intrusion path data record in detail the operation steps of the intruder in the first intrusion data area. For example, an intruder first attempts to exploit a vulnerability to attack, then attempts to escalate permissions, and finally steals data. These data form an intrusion execution operation sequence, namely the first intrusion path data.
[0024] Step S120: determining a plurality of basic intrusion knowledge points based on the first intrusion path data.
[0025] In this embodiment, the server starts to analyze the first intrusion path data, summarizes and concludes these operation steps, and forms multiple basic intrusion knowledge points. For example, the server finds that there are behaviors of using specific vulnerabilities to conduct preliminary attacks in multiple intrusion events, so this behavior is refined into a basic intrusion knowledge point; similarly, privilege escalation and data theft are also refined into basic intrusion knowledge points.
[0026] Step S130, obtaining second intrusion path data of the target intrusion detection event in the security protection platform, wherein the second intrusion path data represents an intrusion execution operation sequence of the corresponding target intrusion detection event in the second intrusion data area, wherein the target intrusion detection event is an intrusion detection event in which a variation occurs in the intrusion execution operation sequence in the second intrusion data area, and the second intrusion data area is an intrusion data area after a business node of the first intrusion data area.
[0027] The server then obtains the second intrusion path data of the target intrusion detection event from the security protection platform. These second intrusion path data reflect that the intruder's operation steps have changed in the second intrusion data area. For example, the intruder may try a new vulnerability exploit or use a different privilege escalation technique. Importantly, the second intrusion data area is the area after the business node of the first intrusion data area, which means that these changes may occur after the intruder has penetrated deep into the system.
[0028] Step S140: determining a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area based on the second intrusion path data and the basic intrusion knowledge point.
[0029] The server compares the second intrusion path data with the basic intrusion knowledge points extracted before. In this process, the server discovers some new intrusion methods that are different from the basic intrusion knowledge points, and these new intrusion methods are identified as target intrusion knowledge points. For example, the server may discover a new method of privilege escalation that has not appeared in previous intrusion incidents and is therefore regarded as a new knowledge point.
[0030] Step S150, determining an event intrusion derived anomaly label of the target intrusion detection event based on the prior intrusion derived anomaly label corresponding to the target intrusion knowledge point.
[0031] Finally, the server determines the event intrusion-derived anomaly labels of the target intrusion detection events based on the prior intrusion-derived anomaly labels corresponding to these target intrusion knowledge points. For example, if a new privilege escalation method is marked as a high-risk behavior, then the target intrusion detection events using this method will also be marked as high-risk. In this way, the security team can quickly identify which intrusion events use new and potentially more threatening attack methods, and take corresponding defensive measures.
[0032] Based on the above steps, by obtaining the first intrusion path data of the intrusion detection event in the security protection platform, and determining multiple basic intrusion knowledge points based on these first intrusion path data, by obtaining the second intrusion path data of the target intrusion detection event, and combining the basic intrusion knowledge points, it is possible to accurately identify the target intrusion knowledge points that are different from the basic intrusion knowledge points in the second intrusion data area, and determine the event intrusion derived anomaly labels of the target intrusion detection event based on the prior intrusion derived anomaly labels corresponding to the target intrusion knowledge points, which not only depends on the current intrusion behavior characteristics, but also combines historical information and prior knowledge, thereby improving the accuracy of label determination. Through accurate identification of intrusion behavior and label determination, it can help the security protection platform to more effectively respond to various network security threats and improve the overall security protection capabilities of the system.
[0033] In a possible implementation, the determining, based on the second intrusion path data and the basic intrusion knowledge point, a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area includes: Based on the second intrusion path data, second intrusion description vector sequences corresponding to the target intrusion detection events are determined.
[0034] The deviation between the second intrusion description vector sequence and the intrusion description vector sequence corresponding to the basic intrusion knowledge point is determined, and the second intrusion description vector sequence whose deviation is greater than a preset deviation is used as a candidate intrusion description vector sequence.
[0035] At least one second intrusion description group is generated according to grouping the candidate intrusion description vector sequence.
[0036] The target intrusion knowledge point is determined based on a candidate intrusion description vector sequence covered by the second intrusion description group.
[0037] In this embodiment, the server first processes the acquired second intrusion path data, which records in detail the operation steps of the target intrusion detection event in the second intrusion data area. The server converts these operation steps into a mathematical vector form, namely, a second intrusion description vector sequence. Each vector represents a specific operation step or behavior feature. For example, a vector may represent the behavior of "trying to access a specific system file", which contains the type, time, frequency and other characteristic information of the behavior. In this way, the server generates a unique second intrusion description vector sequence for each target intrusion detection event.
[0038] The server then compares the second intrusion description vector sequence with the intrusion description vector sequence corresponding to the basic intrusion knowledge point. Here, the intrusion description vector sequence of the basic intrusion knowledge point is extracted in advance based on a large amount of historical intrusion data and represents a common intrusion behavior pattern. The server measures the difference between them by calculating the deviation between each second intrusion description vector sequence and the corresponding sequence of the basic intrusion knowledge point. The deviation can be a numerical indicator that combines multiple feature differences. When the deviation of a second intrusion description vector sequence exceeds a preset threshold (i.e., the preset deviation), the server marks it as a candidate intrusion description vector sequence, which means that it may represent a new intrusion behavior that is different from the known basic intrusion knowledge point.
[0039] The server further processes the intrusion description vector sequences marked as candidates. Since these sequences may have some common features or behavior patterns, the server uses a clustering algorithm or other grouping techniques to group them. The purpose of grouping is to group candidate intrusion description vector sequences that exhibit similar behaviors together to form one or more second intrusion description groups. The sequences within each group have a high similarity in intrusion behavior, while the sequences between different groups are quite different.
[0040] Finally, the server determines the target intrusion knowledge point based on the formed second intrusion description group. Each group represents a new intrusion behavior pattern that is different from the basic intrusion knowledge point. The server analyzes the candidate intrusion description vector sequence in each group, extracts their common features and key behaviors, and thus forms a new intrusion knowledge point, namely the target intrusion knowledge point. These target intrusion knowledge points provide the security protection platform with valuable information about new intrusion behaviors, which helps to update defense strategies in a timely manner and improve the security protection capabilities of the system.
[0041] In a possible implementation, determining the event intrusion derived anomaly label of the target intrusion detection event based on the prior intrusion derived anomaly label corresponding to the target intrusion knowledge point includes: Based on the second intrusion path data corresponding to the second intrusion description group, a priori intrusion-derived anomaly label of the target intrusion knowledge point is determined.
[0042] If a key intelligence intrusion knowledge point whose prior intrusion derived anomaly label is an intelligence trigger label is determined from the target intrusion knowledge point, the target intrusion detection event associated with the second intrusion description group corresponding to the key intelligence intrusion knowledge point is determined as a key intelligence intrusion detection event.
[0043] In a possible implementation manner, the step of grouping the candidate intrusion description vector sequence to generate at least one second intrusion description group includes: The candidate intrusion description vector sequences are subjected to balanced extraction to generate intrusion description vector extraction sequences, wherein the number of the intrusion description vector extraction sequences is less than the number of the candidate intrusion description vector sequences.
[0044] The intrusion description vector extraction sequence is grouped to generate at least one second intrusion description group.
[0045] In this embodiment, after determining the second intrusion description group, the server further analyzes the second intrusion path data corresponding to each group. These second intrusion path data provide the server with detailed context about the intrusion behavior within the group, including the source, target, tools and techniques used, etc. of the attack. Based on this information, the server can assign a priori intrusion-derived anomaly label to each target intrusion knowledge point. These priori intrusion-derived anomaly labels represent different threat levels, attack types, or potential subsequent behaviors. For example, a target intrusion knowledge point may be marked as "high-risk data leakage", indicating that the intrusion behavior is likely to lead to the leakage of sensitive information.
[0046] After assigning the prior intrusion-derived anomaly labels, the server will pay special attention to the target intrusion knowledge points that are marked as intelligence trigger labels. These knowledge points may represent new attack methods, unknown threat sources, or intrusion behaviors with special significance. Once such a key intelligence intrusion knowledge point is identified, the server will trace back to the second intrusion description group associated with it and mark all target intrusion detection events associated with the group as key intelligence intrusion detection events. Doing so helps the security team quickly identify and respond to new attacks that may have a significant impact on system security.
[0047] When processing a large number of candidate intrusion description vector sequences, in order to reduce the complexity of calculation and improve efficiency, the server may adopt a balanced extraction method. This means that the server will evenly select a portion of representative sequences from the candidate intrusion description vector sequences to form an intrusion description vector extraction sequence. This extraction process is balanced, aiming to ensure that the extracted sequence can fully reflect the diversity and characteristics of the original candidate sequence, while reducing data redundancy and computational complexity.
[0048] After completing the balanced extraction, the server will use the extracted intrusion description vector extraction sequences to group them. This grouping process may use a clustering algorithm or other similarity measurement method to cluster extraction sequences with similar characteristics to form a second intrusion description group. Since the extraction sequences are selected in a balanced manner, these groups can more accurately represent the clustering characteristics of the original candidate intrusion description vector sequences. In this way, the server can more effectively identify new intrusion behavior patterns and provide targeted guidance for subsequent security defense.
[0049] In a possible implementation manner, the candidate vector sequence is one of the grouped intrusion description vector sequences corresponding to the basic intrusion knowledge points, and the step of determining the preset deviation includes: A candidate deviation between the second intrusion description vector sequence and the candidate vector sequence is determined.
[0050] Based on a descending set of threshold values, the number of sequences of the second intrusion description vector sequences whose candidate deviations meet the candidate threshold values in the descending set is determined in sequence.
[0051] If the difference in the number of sequences between the number of sequences at the xth candidate threshold value and the number of sequences at the x+1th candidate threshold value meets the set requirement, the preset deviation is determined based on the xth candidate threshold value.
[0052] In a possible implementation manner, determining a plurality of basic intrusion knowledge points based on the first intrusion path data includes: Based on the first intrusion path data, first intrusion description vector sequences corresponding to the intrusion detection events in the security protection platform are determined.
[0053] A plurality of first intrusion description groups are generated by grouping the first intrusion description vector sequence.
[0054] Based on the first intrusion description vector sequence covered by the first intrusion description group, a plurality of basic intrusion knowledge points respectively corresponding to the plurality of first intrusion description groups are determined.
[0055] In this embodiment, the server first selects a candidate vector sequence, which is selected from the grouped intrusion description vector sequence corresponding to the basic intrusion knowledge point. Then, the server calculates the candidate deviation between the second intrusion description vector sequence and the candidate vector sequence. The candidate deviation is a numerical value used to quantify the degree of difference between two vector sequences. The server may use various algorithms to calculate this deviation, such as calculating the Euclidean distance or cosine similarity between the two sequences. The calculated candidate deviation will be used in subsequent steps.
[0056] Furthermore, the server has a descending set of threshold values, which are pre-set and used to divide different deviation intervals. The server will judge whether the candidate deviation of each second intrusion description vector sequence meets the current threshold value according to the threshold values in this descending set. For each threshold value, the server will count the number of second intrusion description vector sequences that meet the threshold value. This process is to find out the distribution of second intrusion description vector sequences in different deviation intervals.
[0057] After counting the number of second intrusion description vector sequences corresponding to each threshold value, the server will check whether the difference in the number of sequences corresponding to two adjacent threshold values (the xth and x+1th) meets a set requirement. This set requirement may be a specific numerical difference threshold, which is used to determine in which deviation interval the change in the number of sequences is most significant. Once a sequence number difference that meets the set requirement is found, the server will use the xth candidate threshold value as the preset deviation. This preset deviation will serve as an important judgment basis in subsequent intrusion detection.
[0058] In another possible implementation, the server will first process the first intrusion path data. These data record the relevant information of the historical intrusion detection events in the security protection platform. The server will convert these historical intrusion detection events into first intrusion description vector sequences, each of which represents the characteristics and behavior patterns of a specific intrusion detection event. This process is similar to the generation process of the second intrusion description vector sequence mentioned above, but here it is based on historical data.
[0059] After generating the first intrusion description vector sequences, the server will group these sequences. The purpose of grouping is to group sequences with similar characteristics and behavior patterns together to form multiple first intrusion description groups. This process may be implemented using clustering algorithms or other similarity measurement methods. Each group represents a specific intrusion behavior type or pattern.
[0060] Finally, the server will determine the corresponding basic intrusion knowledge points based on the first intrusion description vector sequence covered by each first intrusion description group. These basic intrusion knowledge points are a summary and conclusion of the intrusion behavior in the group, including the key characteristics and behavior patterns of this type of intrusion. These basic intrusion knowledge points will serve as an important reference for subsequent intrusion detection, helping the server to quickly identify and respond to potential intrusion threats.
[0061] In a possible implementation manner, determining a plurality of basic intrusion knowledge points respectively corresponding to the plurality of first intrusion description groups based on the first intrusion description vector sequence covered by the first intrusion description group includes: Extraction is performed from a first intrusion description vector sequence covered by the first intrusion description group, and a group intrusion description vector sequence is determined based on the extracted first intrusion description vector sequence.
[0062] The group intrusion description vector sequences corresponding to the plurality of first intrusion description groups are determined as the intrusion description vector sequences of the plurality of basic intrusion knowledge points.
[0063] In a possible implementation manner, the regional service span corresponding to the second intrusion data area is equal to the regional service span corresponding to the first intrusion data area.
[0064] In a possible implementation manner, the candidate intrusion knowledge point is the basic intrusion knowledge point or the target intrusion knowledge point, and in the process of determining, based on the second intrusion path data and the basic intrusion knowledge point, a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area, the method further includes: If the number of pairings between the candidate intrusion knowledge point and the second intrusion path data is less than a set number, the candidate intrusion knowledge point is removed.
[0065] In this embodiment, after completing the grouping of the first intrusion description vector sequence, the server will further extract a representative first intrusion description vector sequence from each first intrusion description group. This process may involve statistical analysis of the vector sequences in each group to find out the sequences that best represent the characteristics of the group. Once these representative sequences are determined, they are called grouped intrusion description vector sequences. These sequences will serve as an important basis for determining basic intrusion knowledge points in the subsequent process.
[0066] After extracting the grouped intrusion description vector sequences, the server will associate these sequences with the corresponding first intrusion description groups. Each grouped intrusion description vector sequence is regarded as the core description of a basic intrusion knowledge point. These basic intrusion knowledge points are abstractions and summaries of various known intrusion behaviors, and they contain the key features and patterns of intrusion behaviors. By corresponding the grouped intrusion description vector sequences with the basic intrusion knowledge points, the server can establish a comprehensive intrusion knowledge system to provide strong support for subsequent security protection.
[0067] In this possible implementation, the regional business span corresponding to the second intrusion data area is equal to the regional business span corresponding to the first intrusion data area. This means that the business scope, system architecture, data flow, etc. involved in the two data areas are similar or consistent. Such a setting helps to compare and analyze intrusion behaviors in the same business context, thereby improving the accuracy and effectiveness of intrusion detection.
[0068] If the number of times a candidate intrusion knowledge point (whether it is a basic intrusion knowledge point or a target intrusion knowledge point) is matched with the second intrusion path data is less than a set number threshold, then the candidate intrusion knowledge point will be removed. The purpose of this step is to filter out those uncommon or occasional intrusion behaviors and ensure that the final intrusion knowledge point is representative and stable. In this way, the server can focus on those intrusion behaviors that really pose a threat to system security and improve the efficiency and accuracy of security protection.
[0069] In the above embodiments, the artificial intelligence-based anomaly identification system for executing the above method embodiments has at least one processor, a control module (chip set) coupled to at least one of the (at least one) processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one input / output device coupled to the control module, and a network interface coupled to the control module.
[0070] The processor may include at least one single-core or multi-core processor, and the processor may include any combination of a general-purpose processor or a dedicated processor (such as a graphics processor, an application processor, a baseband processor, etc.). For some alternative implementations, the artificial intelligence-based anomaly recognition system can be used as an artificial intelligence-based anomaly recognition system device such as a gateway described in the embodiments of the present application.
[0071] For some alternative embodiments, the artificial intelligence-based anomaly identification system may include at least one computer-readable medium (e.g., a memory or NVM / storage device) having instructions and at least one processor integrated with the at least one computer-readable medium and configured to execute the instructions to implement a module to perform the actions described in the present disclosure.
[0072] For one embodiment, the control module may include any suitable interface controller to provide any suitable interface to at least one of the processor(s) and / or any suitable device or component in communication with the control module.
[0073] The control module may include a memory controller module to provide an interface to the memory. The memory controller module may be a hardware module, a software module, and / or a firmware module.
[0074] The memory may be used, for example, to load and store data and / or instructions for an artificial intelligence-based anomaly recognition system. For one embodiment, the memory may include any suitable volatile memory, such as a suitable DRAM.
[0075] For one embodiment, the control module may include at least one input / output controller to provide an interface to the NVM / storage device and the (at least one) input / output device.
[0076] For example, the NVM / storage device may be used to store data and / or instructions. The NVM / storage device may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one compact disk (CD) drive, and / or at least one digital versatile disk (DVD) drive).
[0077] The NVM / storage device may include storage resources that are physically part of the device on which the artificial intelligence-based anomaly identification system is installed, or it may be accessible to the device without being part of the device. For example, the NVM / storage device may be accessed via (at least one) input / output device over a network.
[0078] The (at least one) input / output device may provide an interface for the artificial intelligence-based anomaly identification system to communicate with any other appropriate device, and the input / output device may include a communication component, a phonetic component, a sensor component, etc. The network interface may provide an interface for the artificial intelligence-based anomaly identification system to predict network communications based on at least one user preference, and the artificial intelligence-based anomaly identification system may wirelessly communicate with at least one component of a wireless network based on any of at least one wireless network standard and / or protocol, such as accessing a wireless network based on a communication standard.
[0079] For one embodiment, at least one of the (at least one) processor may be loaded together with the logic of at least one controller of the control module (e.g., a memory controller module). For one embodiment, at least one of the (at least one) processor may be loaded together with the logic of at least one controller of the control module to form a system level load. For one embodiment, at least one of the (at least one) processor may be integrated on the same die with the logic of at least one controller of the control module. For one embodiment, at least one of the (at least one) processor may be integrated on the same die with the logic of at least one controller of the control module to form a system on chip (SoC).
[0080] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of the present application. At the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
[0081] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the artificial intelligence-based anomaly identification method described in the aforementioned embodiment.
[0082] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps in the artificial intelligence-based anomaly identification method described in the aforementioned embodiment.
[0083] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple user preference prediction network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.
[0084] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other medium that can be used to be computer-readable with or store data. Finally, it should be noted that what is disclosed above is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. 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. An abnormality identification method based on artificial intelligence, characterized in that: The method comprises: Acquire first intrusion path data of an intrusion detection event in the security protection platform, wherein the first intrusion path data represents an intrusion execution operation sequence of the corresponding intrusion detection event in the first intrusion data area; Determine a plurality of basic intrusion knowledge points based on the first intrusion path data; Acquire second intrusion path data of a target intrusion detection event in the security protection platform, wherein the second intrusion path data represents an intrusion execution operation sequence of the corresponding target intrusion detection event in a second intrusion data area, wherein the target intrusion detection event is an intrusion detection event in which a variation occurs in the intrusion execution operation sequence in the second intrusion data area, and the second intrusion data area is an intrusion data area after a service node of the first intrusion data area; Based on the second intrusion path data and the basic intrusion knowledge point, determining a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area; Based on the prior intrusion-derived anomaly label corresponding to the target intrusion knowledge point, an event intrusion-derived anomaly label of the target intrusion detection event is determined.
2. The artificial intelligence-based anomaly identification method according to claim 1, characterized in that: The determining, based on the second intrusion path data and the basic intrusion knowledge point, a target intrusion knowledge point of the target intrusion detection event that is different from the basic intrusion knowledge point in the second intrusion data area includes: Determine, based on the second intrusion path data, second intrusion description vector sequences corresponding to the target intrusion detection events respectively; Determine the deviation between the second intrusion description vector sequence and the intrusion description vector sequence corresponding to the basic intrusion knowledge point, and use the second intrusion description vector sequence with a deviation greater than a preset deviation as a candidate intrusion description vector sequence; generating at least one second intrusion description group according to grouping the candidate intrusion description vector sequence; The target intrusion knowledge point is determined based on a candidate intrusion description vector sequence covered by the second intrusion description group.
3. The artificial intelligence-based anomaly identification method according to claim 2, characterized in that: The determining, based on the prior intrusion derived anomaly label corresponding to the target intrusion knowledge point, the event intrusion derived anomaly label of the target intrusion detection event comprises: Determining a priori intrusion-derived anomaly label of the target intrusion knowledge point based on second intrusion path data corresponding to the second intrusion description group; If a key intelligence intrusion knowledge point whose prior intrusion derived abnormal label is an intelligence trigger label is determined from the target intrusion knowledge point, the target intrusion detection event associated with the second intrusion description group corresponding to the key intelligence intrusion knowledge point is determined as a key intelligence intrusion detection event.
4. The artificial intelligence-based anomaly identification method according to claim 2, characterized in that: The method of grouping the candidate intrusion description vector sequence according to the method to generate at least one second intrusion description group comprises: Performing balanced extraction on the candidate intrusion description vector sequences to generate intrusion description vector extraction sequences, wherein the number of the intrusion description vector extraction sequences is less than the number of the candidate intrusion description vector sequences; The intrusion description vector extraction sequence is grouped to generate at least one second intrusion description group.
5. The artificial intelligence-based anomaly identification method according to claim 2, characterized in that: The candidate vector sequence is one of the grouped intrusion description vector sequences corresponding to the basic intrusion knowledge point, and the step of determining the preset deviation degree includes: determining a candidate deviation between the second intrusion description vector sequence and the candidate vector sequence; Based on a descending set of threshold values, sequentially determining the number of sequences of the second intrusion description vector sequences whose candidate deviations meet the candidate threshold values in the descending set; If the difference in the number of sequences between the number of sequences at the xth candidate threshold value and the number of sequences at the x+1th candidate threshold value meets the set requirement, the preset deviation is determined based on the xth candidate threshold value.
6. The artificial intelligence-based anomaly identification method according to claim 1, characterized in that: The determining a plurality of basic intrusion knowledge points based on the first intrusion path data includes: Determine, based on the first intrusion path data, first intrusion description vector sequences corresponding to the intrusion detection events in the security protection platform; generating a plurality of first intrusion description groups according to grouping the first intrusion description vector sequence; Based on the first intrusion description vector sequence covered by the first intrusion description group, a plurality of basic intrusion knowledge points respectively corresponding to the plurality of first intrusion description groups are determined.
7. The artificial intelligence-based anomaly identification method according to claim 6, characterized in that: The determining, based on the first intrusion description vector sequence covered by the first intrusion description group, a plurality of basic intrusion knowledge points respectively corresponding to the plurality of first intrusion description groups comprises: Extracting from a first intrusion description vector sequence covered by the first intrusion description group, and determining a group intrusion description vector sequence based on the extracted first intrusion description vector sequence; The group intrusion description vector sequences corresponding to the plurality of first intrusion description groups are determined as the intrusion description vector sequences of the plurality of basic intrusion knowledge points.
8. The artificial intelligence-based anomaly identification method according to any one of claims 1 to 7, characterized in that: The regional service span corresponding to the second intrusion data area is equal to the regional service span corresponding to the first intrusion data area.
9. The artificial intelligence-based anomaly identification method according to any one of claims 1 to 7, characterized in that: The candidate intrusion knowledge point is the basic intrusion knowledge point or the target intrusion knowledge point. In the process of determining the target intrusion knowledge point that is different from the basic intrusion knowledge point in the second intrusion data area of the target intrusion detection event based on the second intrusion path data and the basic intrusion knowledge point, the method also includes: if the number of pairings between the candidate intrusion knowledge point and the second intrusion path data is less than a set number, the candidate intrusion knowledge point is removed.
10. An abnormality recognition system based on artificial intelligence, characterized in that: It includes a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the artificial intelligence-based anomaly recognition method described in any one of claims 1 to 9 is implemented.
