CN113536302A - Interface caller safety rating method based on deep learning - Google Patents

Interface caller safety rating method based on deep learning Download PDF

Info

Publication number
CN113536302A
CN113536302A CN202110841579.4A CN202110841579A CN113536302A CN 113536302 A CN113536302 A CN 113536302A CN 202110841579 A CN202110841579 A CN 202110841579A CN 113536302 A CN113536302 A CN 113536302A
Authority
CN
China
Prior art keywords
vector
model
safety
user
feature vector
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110841579.4A
Other languages
Chinese (zh)
Inventor
钱丰
王佳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Institute of Computer Technology and Applications
Original Assignee
Beijing Institute of Computer Technology and Applications
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Institute of Computer Technology and Applications filed Critical Beijing Institute of Computer Technology and Applications
Priority to CN202110841579.4A priority Critical patent/CN113536302A/en
Publication of CN113536302A publication Critical patent/CN113536302A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/50—Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems
    • G06F21/55—Detecting local intrusion or implementing counter-measures
    • G06F21/554—Detecting local intrusion or implementing counter-measures involving event detection and direct action
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Computer Security & Cryptography (AREA)
  • Mathematical Physics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Molecular Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Hardware Design (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to an interface caller safety rating method based on deep learning, and belongs to the field of information safety. The method includes the steps that an interface caller characteristic vector is constructed by collecting behavior characteristics of an interface caller; inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a depth learning technology; and inputting the analysis discrimination vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the interface caller safety through the Logistic multi-classification regression analysis model. The invention is applied to the safety rating of the interface caller, thereby enhancing the safety monitoring of the interface calling, increasing the safety of the interface calling and enhancing the data safety.

Description

Interface caller safety rating method based on deep learning
Technical Field
The invention belongs to the field of information security, and particularly relates to an interface caller security rating method based on deep learning.
Background
Interface caller security rating, i.e. rating of security for interface caller identity, is a process of analyzing, processing, and classifying users with different security levels. With the popularization of network interface calling technology and crawler technology, the existing system often receives a large amount of interface calling requests. The security of the user behind these requests is not necessarily guaranteed, so the interface caller security rating technique is gaining increasing attention. The interface caller safety rating is an important work in system safety protection, is mainly used for the aspects of system protection, intrusion detection, user safety identification and the like, and plays a vital role in system safety and system data safety.
With the increasing complexity of the identity of the user called by the online request interface, the conventional method of manually writing the user by a rule and then judging and classifying the user by using conditions cannot meet the increasing requirement on the security. Firstly, the method of writing rules by hand and then judging and classifying by using conditions usually hard codes the rules and writes them into a system, so that the rules cannot be modified at will, which often leads to the situation that the manual rules are very lagged behind the reality. In addition, the method of manually writing the rule and then judging and classifying the use conditions usually requires a lot of manual work, which causes great strain on the work efficiency.
After the interface caller safety rating method based on the deep learning technology and the Logistic multi-classification regression analysis model disclosed by the invention is used, the system can conveniently and quickly modify the rules by changing the characteristic vector table. In addition, the structure of the depth model does not need to be greatly changed according to the rule, so that the model is favorable to advance with time and is kept closely synchronous with the reality.
Therefore, there is an urgent need for a method for further utilizing the security rating capability of deep learning interface callers to solve the problems of difficulty in writing manual rules and slow update under the condition of satisfying effective rating of the security of the interface callers. The present invention has been made in a formal sense to meet this real-world need.
Disclosure of Invention
Technical problem to be solved
The invention provides a method for grading interface caller safety based on deep learning, and aims to solve the problems that manual rules are difficult to write and updating is slow under the condition that interface caller safety is effectively graded.
(II) technical scheme
In order to solve the technical problem, the invention provides an interface caller security rating method based on deep learning, which comprises the following steps:
s1, constructing an interface caller feature vector by collecting the behavior features of the interface caller;
s2, inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a deep learning technology;
and S3, inputting the analysis and judgment vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the safety of the interface caller through the Logistic multi-classification regression analysis model.
Further, the step S1 specifically includes: a feature vector table is established, the behavior features of specific callers are encoded into feature vectors through the feature vector table, and the behavior features of each user are encoded into a feature vector { x) with fixed length through the feature vector table1,x2,x3..,xnWhere n is the dimension of the feature vector.
Further, each feature vector should be given a discrete integer feature value.
Further, the behavior characteristics of the caller comprise the number of times of access per second, the number of times of access today, the number of times of access this month, the number of accessed data sets, user registration time, user permission type, the system category to which the user belongs, and the user download data volume.
Further, the construction of the feature vector table needs to consider several factors: the unit has the user behavior data, whether the data can sufficiently reveal the safety of the user, and whether the data has the distinguishing capability from the technical point of view in the model building process.
Further, the step S2 specifically includes: feature vector x for each user1,x2,x3..,xnAn input to a fully connected deep learning model will result in an m-dimensionalOutput vector { a1,a2,a3,…amAfter the output vector is normalized by softmax, the vector { b can be output1,b2,b3,…bmIn which b isi∈[0,1]The output vector is the output analysis discrimination vector.
Further, the deep learning model has a fully connected node structure.
Further, the dimension structure of the depth model is n-l1-l2-l3M, where n is the feature vector { x }1,x2,x3..,xnDimension n, m of the device is dimension m, l of the output analysis discriminant direction1,l2,l3Is the middle dimension of the fully connected depth model.
Further, the step S3 specifically includes: output analysis discriminant vector b for individual user1,b2,b3,…bmInputting the data into a Logistic multi-classification regression analysis model, and judging the input analysis discrimination vector { b ] by the Logistic multi-classification regression analysis model1,b2,b3,…bmGive class c, where c ∈ [0,1,2,3 … j]Where j is the classification for the security level in the training model.
Further, the classification of the security level includes: very safe, the corresponding number is 1; safety, the corresponding number is 2; fuzzy, corresponding number is 3; unsafe, and the corresponding number is 4; it is very unsafe, corresponding to number 5.
(III) advantageous effects
Compared with the existing interface caller safety classification rating method based on rules, the interface caller safety rating method based on deep learning has the advantages that in the interface caller safety rating, the problems of poor timeliness and low efficiency of traditional manual rule coding are solved on the aspect of user-level safety analysis by utilizing a deep learning technology and a Logistic multi-classification regression analysis model, and the safety classification ratings of the existing users and the newly-added users can be judged and predicted through the past data. The invention can provide caller safety classification rating and improve classification performance and accuracy rate on the premise of ensuring correct operation data record of the interface caller. Therefore, the invention will play an important role in interface caller security rating.
Drawings
FIG. 1 is a flow diagram illustrating a method for interface caller security rating based on deep learning technology and Logistic multi-classification regression analysis model according to the present invention.
Detailed Description
In order to make the objects, contents and advantages of the present invention clearer, the following detailed description of the embodiments of the present invention will be made in conjunction with the accompanying drawings and examples.
The invention is applied to the safety rating of the interface caller, thereby enhancing the safety monitoring of the interface calling, increasing the safety of the interface calling and enhancing the data safety. By utilizing a deep learning technology and a Logistic multi-classification regression analysis model, the problems of poor timeliness and low efficiency of traditional manual rule coding are solved on the aspect of user-level safety analysis, and the safety classification grades of the existing and newly-added users can be judged and predicted through the previous data.
The invention discloses an interface caller safety rating method based on a deep learning technology and a Logistic multi-classification regression analysis model, which comprises the following steps: (1) constructing an interface caller characteristic vector by collecting the behavior characteristics of an interface caller; (2) inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a deep learning technology; (3) and inputting the analysis discrimination vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the interface caller safety through the Logistic multi-classification regression analysis model.
The invention is applied to the safety rating of the interface caller, thereby enhancing the safety monitoring of the interface calling, increasing the safety of the interface calling and enhancing the data safety. By utilizing a deep learning technology and a Logistic multi-classification regression analysis model, the problems of poor timeliness and low efficiency of traditional manual rule coding are solved on the aspect of user-level safety analysis, and the safety classification grades of the existing and newly-added users can be judged and predicted through the previous data. The invention has wide application in the field of interface call safety monitoring.
In order to achieve the above object, the present invention provides an interface caller security rating method based on deep learning technology and Logistic multi-classification regression analysis model, the interface caller security rating method includes the steps of:
s1, constructing an interface caller feature vector by collecting the behavior features of the interface caller;
s2, inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a deep learning technology;
and S3, inputting the analysis and judgment vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the safety of the interface caller through the Logistic multi-classification regression analysis model.
The invention discloses an interface caller safety rating method based on a deep learning technology and a Logistic multi-classification regression analysis model, and as shown in the figure, the interface caller safety rating step of the method comprises the following steps:
s1, constructing an interface caller feature vector by collecting the behavior features of the interface caller;
in the specific implementation, a feature vector table is first prepared, and the behavior features of a specific caller are encoded into a feature vector through the feature vector table. Each eigenvector should be given a discrete integer eigenvalue. Finally, the behavior characteristics of each user are coded into a characteristic vector { x with fixed length by a characteristic vector table1,x2,x3..,xnWhere n is the dimension of the feature vector.
For example, the following feature vector table is now given:
TABLE 1 feature vector Table
Feature(s) Value taking
Number of accesses per second Integer number of
Number of visits today Integer number of
Number of visits in this month Integer number of
Number of data sets accessed Integer number of
User registration time (day) Integer number of
Type of user's right Categorizing representatives by integer values
System class to which user belongs Categorizing representatives by integer values
User download data volume (MB) Integer, less than 1MB, calculated as 1MB
Etc. …
The behavior of a certain user is: the number of accesses per second is 3, the number of accesses per day is 52, the number of accesses per month is 2042, the number of accessed data sets is 32, the user registration time (day) is 105 days, the user permission type is secret (the category of the secret is 1), the system category to which the user belongs is an enterprise user (the category of the enterprise user is 3), and the user download data volume (MB) is 200 MB. The feature vector constructed for this user is {3,52,2042,32,105,1,3,200}
It should be noted that the above feature vectors are only examples of this patent, and in practice, the final feature vector table is constructed by considering several factors:
what user behavior data the unit in which it is located possesses
Whether the data can sufficiently reveal the security of the user
Whether this data has a resolving power in the model building process from a technical point of view.
S2, inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a deep learning technology;
in specific implementation, firstly, the feature vector { x ] of each user is1,x2,x3..,xnThe input is to a standard fully connected deep learning model. The deep learning model has a fully connected node structure. The dimension structure of the depth model is n-l1-l2-l3M, where n is the feature vector { x }1,x2,x3..,xnDimension n, m of the device is dimension m, l of the output analysis discriminant direction1,l2,l3The intermediate dimension of the full-connection depth model can be adjusted according to the actual model expression. Feature vector of user { x1,x2,x3..,xnAfter the input of the vector is calculated by the deep learning model, an m-dimensional output vector { a } is obtained1,a2,a3,…am}. After the output vector is normalized by softmax, the vector { b can be output1,b2,b3,…bmIn which b isi∈[0,1]The output vector is the output analysis discrimination vector.
For example, the feature vector constructed by the user is {3,52,2042,32,105,1,3,200}, and the dimension of the feature vector is 8. A certain fully-connected deep learning model has a fully-connected node structure. The dimension structure of the depth model is 8-50-100-50-10. The feature vectors {3,52,2042,32,105,1,3,200} of the user are input into the depth model for calculation. And outputting a 10-dimensional output vector {1.5,20.1,0.54,23.12,8.23,101.23,34.45,0.023,32 and 28}, wherein the output vector is normalized by softmax to output a classification discrimination vector {0.0807,0.854,0.134,0.986,0.535,0.211,0.885,0.949,0.323 and 0.454 }. This vector will be the primary discrimination vector for the next regression. (the number in the vector is only schematic)
And S3, inputting the analysis and judgment vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the safety of the interface caller through the Logistic multi-classification regression analysis model.
In specific implementation, the output of a single user is analyzed to determine the vector { b }1,b2,b3,…bmInputting the data into a Logistic multi-classification regression analysis model known in the art, and judging the vector b according to the input analysis by the Logistic multi-classification regression analysis model1,b2,b3,…bmGive class c, where c ∈ [0,1,2,3 … j]Where j is the classification for the security level in the training model.
For example: inputting the classification discriminant vectors {0.0807,0.854,0.134,0.986,0.535,0.211,0.885,0.949,0.323 and 0.454} of the single user into a Logistic multi-classification regression analysis model, and classifying the input analysis discriminant vectors by the Logistic multi-classification regression analysis model to give a classification c which is 3.
The specific classification table may be a table (the specific classification table needs to be redesigned by the relevant user according to the actual situation of the business, and the classification table is just an example):
TABLE 2 Classification of tables
Level of security Classification number
Is very safe, advised by 1
Safety, re-assessment after 1 day of recommendation 2
Fuzzy, suggesting manual review 3
Is not safe and refuses 4
Very unsafe and immediately alarm 5
Etc. of …
The classification of the user is classification c-3, fuzzy, suggesting manual review.
The invention relates to an interface caller safety rating method based on a deep learning technology and a Logistic multi-classification regression analysis model, which comprises 3 steps:
(1) constructing an interface caller characteristic vector by collecting the behavior characteristics of an interface caller;
(2) inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a deep learning technology;
(3) and inputting the analysis discrimination vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the interface caller safety through the Logistic multi-classification regression analysis model.
Further, the step 1 comprises:
first, a feature vector table should be prepared and storedThe behavior characteristics of the fixed caller are encoded into a feature vector through a feature vector table. Each feature vector should be given a discrete integer feature value. Finally, the behavior characteristics of each user are coded into a characteristic vector { x with fixed length by a characteristic vector table1,x2,x3..,xnWhere n is the dimension of the feature vector.
Further, the step 2 comprises:
in specific implementation, firstly, the feature vector { x ] of each user is1,x2,x3..,xnThe input is to a standard fully connected deep learning model. The deep learning model has a fully connected node structure. The dimension structure of the depth model is n-l1-l2-l3M, where n is the feature vector { x }1,x2,x3..,xnDimension n, m of the device is dimension m, l of the output analysis discriminant direction1,l2,l3The intermediate dimension of the full-connection depth model can be adjusted according to the actual model expression. Feature vector of user { x1,x2,x3..,xnAfter the input of the vector is calculated by the deep learning model, an m-dimensional output vector { a } is obtained1,a2,a3,…am}. After the output vector is normalized by softmax, the vector { b can be output1,b2,b3,…bmIn which b isi∈[0,1]The output vector is the output analysis discrimination vector.
Further, the step 3 comprises:
output analysis discrimination vector { b) for a single user1,b2,b3,…bmInputting the data into a Logistic multi-classification regression analysis model known in the art, and judging the vector b according to the input analysis by the Logistic multi-classification regression analysis model1,b2,b3,…bmGive class c, where c ∈ [0,1,2,3 … j]Where j is the classification for the security level in the training model.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims (10)

1. A method for interface caller safety rating based on deep learning is characterized by comprising the following steps:
s1, constructing an interface caller feature vector by collecting the behavior features of the interface caller;
s2, inputting the feature vector of the previous step into a depth model, and calculating an output analysis discrimination vector through a deep learning technology;
and S3, inputting the analysis and judgment vector into a Logistic multi-classification regression model, and finally obtaining the classification rating of the safety of the interface caller through the Logistic multi-classification regression analysis model.
2. The method for security rating of an interface caller based on deep learning of claim 1, wherein the step S1 specifically comprises: a feature vector table is established, the behavior features of specific callers are encoded into feature vectors through the feature vector table, and the behavior features of each user are encoded into a feature vector { x) with fixed length through the feature vector table1，x2，x3..，xnWhere n is the dimension of the feature vector.
3. The method of claim 2, wherein each feature vector is given a discrete integer feature value.
4. The method as claimed in claim 2, wherein the behavior characteristics of the caller include number of accesses per second, number of accesses today, number of accesses this month, number of data sets accessed, user registration time, user permission type, system category to which the user belongs, and user download data volume.
5. The method as claimed in claim 2, wherein the feature vector table is constructed by considering several factors: the unit has the user behavior data, whether the data can sufficiently reveal the safety of the user, and whether the data has the distinguishing capability from the technical point of view in the model building process.
6. The method for interface caller security rating based on deep learning of any one of claims 1 to 5, wherein the step S2 specifically comprises: feature vector x for each user1，x2，x3..，xnInputting into a fully connected deep learning model, an m-dimensional output vector { a } will be obtained1，a2，a3，…amAfter the output vector is normalized by softmax, the vector { b can be output1，b2，b3，…bmIn which b isi∈[0，1]The output vector is the output analysis discrimination vector.
7. The method of claim 6, wherein the deep learning model has a fully connected node structure.
8. The method as claimed in claim 7, wherein the dimension structure of the depth model is n-l1-l2-l3M, where n is the feature vector { x }1，x2，x3..，xnDimension n, m of the device is dimension m, l of the output analysis discriminant direction1，l2，l3Is the middle dimension of the fully connected depth model.
9. The method of claim 7 or 8, wherein the interface caller security rating based on deep learningStep S3 specifically includes: output analysis discriminant vector b for individual user1，b2，b3，…bmInputting the data into a Logistic multi-classification regression analysis model, and judging the input analysis discrimination vector { b ] by the Logistic multi-classification regression analysis model1，b2，b3，…bmGive class c, where c ∈ [0,1,2, 3.. j]Where j is the classification for the security level in the training model.
10. The deep learning based interface caller security rating method of claim 9, wherein the classification of the security level comprises: very safe, the corresponding number is 1; safety, the corresponding number is 2; fuzzy, corresponding number is 3; unsafe, and the corresponding number is 4; it is very unsafe, corresponding to number 5.
CN202110841579.4A 2021-07-26 2021-07-26 Interface caller safety rating method based on deep learning Pending CN113536302A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110841579.4A CN113536302A (en) 2021-07-26 2021-07-26 Interface caller safety rating method based on deep learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110841579.4A CN113536302A (en) 2021-07-26 2021-07-26 Interface caller safety rating method based on deep learning

Publications (1)

Publication Number Publication Date
CN113536302A true CN113536302A (en) 2021-10-22

Family

ID=78120714

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110841579.4A Pending CN113536302A (en) 2021-07-26 2021-07-26 Interface caller safety rating method based on deep learning

Country Status (1)

Country Link
CN (1) CN113536302A (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105205396A (en) * 2015-10-15 2015-12-30 上海交通大学 Detecting system for Android malicious code based on deep learning and method thereof
CN107426199A (en) * 2017-07-05 2017-12-01 浙江鹏信信息科技股份有限公司 A kind of method and system of Network anomalous behaviors detection and analysis
US20190166141A1 (en) * 2017-11-30 2019-05-30 Shape Security, Inc. Detection of malicious activity using behavior data
WO2020136497A1 (en) * 2018-12-26 2020-07-02 International Business Machines Corporation Detecting inappropriate activity in the presence of unauthenticated api requests using artificial intelligence
CN111600894A (en) * 2020-05-20 2020-08-28 新华三信息安全技术有限公司 Network attack detection method and device
CN112069485A (en) * 2020-06-12 2020-12-11 完美世界(北京)软件科技发展有限公司 Safety processing method, device and equipment based on user behaviors
US20210160247A1 (en) * 2018-04-02 2021-05-27 Visa International Service Association Real-time entity anomaly detection
CN113010895A (en) * 2020-12-08 2021-06-22 四川大学 Vulnerability hazard assessment index technology based on deep learning

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105205396A (en) * 2015-10-15 2015-12-30 上海交通大学 Detecting system for Android malicious code based on deep learning and method thereof
CN107426199A (en) * 2017-07-05 2017-12-01 浙江鹏信信息科技股份有限公司 A kind of method and system of Network anomalous behaviors detection and analysis
US20190166141A1 (en) * 2017-11-30 2019-05-30 Shape Security, Inc. Detection of malicious activity using behavior data
US20210160247A1 (en) * 2018-04-02 2021-05-27 Visa International Service Association Real-time entity anomaly detection
WO2020136497A1 (en) * 2018-12-26 2020-07-02 International Business Machines Corporation Detecting inappropriate activity in the presence of unauthenticated api requests using artificial intelligence
CN113302609A (en) * 2018-12-26 2021-08-24 国际商业机器公司 Detecting inappropriate activity in the presence of unauthenticated API requests using artificial intelligence
CN111600894A (en) * 2020-05-20 2020-08-28 新华三信息安全技术有限公司 Network attack detection method and device
CN112069485A (en) * 2020-06-12 2020-12-11 完美世界(北京)软件科技发展有限公司 Safety processing method, device and equipment based on user behaviors
CN113010895A (en) * 2020-12-08 2021-06-22 四川大学 Vulnerability hazard assessment index technology based on deep learning

Similar Documents

Publication Publication Date Title
CN109657947B (en) Enterprise industry classification-oriented anomaly detection method
CN111783875A (en) Abnormal user detection method, device, equipment and medium based on cluster analysis
CN109255586B (en) Online personalized recommendation method for e-government affairs handling
CN112418520B (en) Credit card transaction risk prediction method based on federal learning
CN105590055A (en) Method and apparatus for identifying trustworthy user behavior in network interaction system
CN110502638B (en) Enterprise news risk classification method based on target entity
CN112989414B (en) Mobile service data desensitization rule generation method based on width learning
CN112100398B (en) Patent blank prediction method and system
WO2020048056A1 (en) Risk decision method and apparatus
CN112215604A (en) Method and device for identifying information of transaction relationship
CN110210858A (en) A kind of air control guard system design method based on intelligent terminal identification
CN115577152B (en) Online book borrowing management system based on data analysis
CN113902037A (en) Abnormal bank account identification method, system, electronic device and storage medium
CN111191825A (en) User default prediction method and device and electronic equipment
CN108564465A (en) A kind of enterprise credit management method
CN114819967A (en) Data processing method and device, electronic equipment and computer readable storage medium
CN113536302A (en) Interface caller safety rating method based on deep learning
CN115994381A (en) Sensitive data identification method and system for project secret assessment
CN115907968A (en) Wind control rejection inference method and device based on pedestrian credit
CN113177164B (en) Multi-platform collaborative new media content monitoring and management system based on big data
CN112069392B (en) Method and device for preventing and controlling network-related crime, computer equipment and storage medium
CN114036953A (en) Financial field entity and intention identification method based on priori knowledge
CN114338058A (en) Information processing method, device and storage medium
Demir et al. Institutional similarity, firm heterogeneity and export sophistication
CN110766091A (en) Method and system for identifying road loan partner

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination