CN111125539B - CDN harmful information blocking method and system based on artificial intelligence - Google Patents
CDN harmful information blocking method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN111125539B CN111125539B CN201911414631.7A CN201911414631A CN111125539B CN 111125539 B CN111125539 B CN 111125539B CN 201911414631 A CN201911414631 A CN 201911414631A CN 111125539 B CN111125539 B CN 111125539B
- Authority
- CN
- China
- Prior art keywords
- data
- layer
- hidden layer
- request operation
- operation log
- 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.)
- Active
Links
- 230000000903 blocking effect Effects 0.000 title claims abstract description 40
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 25
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000001514 detection method Methods 0.000 claims abstract description 47
- 238000001914 filtration Methods 0.000 claims abstract description 11
- 230000000977 initiatory effect Effects 0.000 claims abstract description 4
- 210000002569 neuron Anatomy 0.000 claims description 38
- 238000012545 processing Methods 0.000 claims description 34
- 238000012544 monitoring process Methods 0.000 claims description 24
- 230000005540 biological transmission Effects 0.000 claims description 16
- 238000012937 correction Methods 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 2
- 238000002955 isolation Methods 0.000 abstract description 3
- 230000006978 adaptation Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
- G06F16/9566—URL specific, e.g. using aliases, detecting broken or misspelled links
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a CDN harmful information blocking method and system based on artificial intelligence, which relate to the technical field of network security and specifically comprise the following steps: acquiring request operation log data of users in all CDN nodes; initiating a corresponding service request according to the request operation log data and acquiring returned data; the method comprises the steps of obtaining a detection result of unstructured data in returned data by adopting an artificial intelligent model, and obtaining the detection result of the structured data in the returned data by keyword filtering and URL filtering; when the detection result is judged to be harmful information, a blocking rule of the request operation log data corresponding to the harmful information is generated and sent to all CDN nodes, and the method and the device can detect unstructured data and perform blocking isolation.
Description
Technical Field
The invention relates to the technical field of network security, in particular to a CDN harmful information blocking method and system based on artificial intelligence.
Background
With the continuous upgrade of hardware devices, the content storage capacity of a CDN (Content Delivery Network ) system is larger and larger, the service types are richer, and the media participating in the system are more complex; unhealthy information such as phishing websites, yellow, terrorism and false information is more and more hidden, and data structures are more and more complex, and traditional filtering based on keywords and URL (uniform resource locator) cannot analyze and process unstructured data such as video and audio, so that the unstructured data cannot be effectively blocked and isolated.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an artificial intelligence-based CDN harmful information blocking method and system, which can finish detection of unstructured data and perform blocking isolation.
In order to achieve the above purpose, the invention adopts the following technical scheme:
in a first aspect, an artificial intelligence based CDN harmful information blocking method is provided, which specifically includes the following steps:
acquiring request operation log data of users in all CDN nodes;
initiating a corresponding service request according to the request operation log data and acquiring returned data;
the method comprises the steps of obtaining a detection result of unstructured data in returned data by adopting an artificial intelligent model, and obtaining the detection result of the structured data in the returned data by keyword filtering and URL filtering;
and when the detection result is judged to be harmful information, generating a blocking rule of the request operation log data corresponding to the harmful information, and sending the blocking rule to all CDN nodes.
On the basis of the technical scheme, the artificial intelligent model adopts an artificial intelligent algorithm to finish detection of unstructured data, and when the artificial intelligent algorithm is adopted to finish detection of unstructured data, the artificial intelligent algorithm corrects the unstructured data through a plurality of known detection results.
On the basis of the technical scheme, the artificial intelligence algorithm comprises an input layer, an implicit layer and an output layer, and corrects the artificial intelligence algorithm through unstructured data of a plurality of known detection results, and specifically comprises the following steps:
initializing a connection weight from an input layer to an implicit layer unit and a connection weight from the implicit layer to an output layer;
sequentially inputting a plurality of data samples with known detection results to obtain the output of an implicit layer and an output layer of each data sample;
obtaining a reverse transmission error from each data sample output layer to the hidden layer and a reverse transmission error from the hidden layer to the input layer according to the output of each data sample hidden layer and the output layer;
and determining a threshold value of the output layer, and obtaining the connection weight between the layers after correction according to the reverse transmission error from the output layer to the hidden layer and the reverse transmission error from the hidden layer to the input layer of each data sample.
Based on the technical scheme, the reverse transmission error from each data sample output layer to the hidden layer and the reverse transmission error from the hidden layer to the input layer are obtained according to the output of each data sample hidden layer and the output layer, and the following formulas are satisfied:
δ k ＝(d k -0 k )f′(net k )
in delta k D, for outputting the feedback error from the layer to the hidden layer k Target value of neurons of the K th layer of hidden layer, 0 k For the output layer, the Kth neuron outputs a value, f' (net k ) Is 0 to k And d k Deviation variance value, delta j For implicit layer to input layer feedback error, f' (net j ) For inputting the target value of the j-th layer neuron and the output value of the j-th layer neuron, omega jk The connection weight of the kth neuron from the hidden layer to the jth layer of the output layer.
On the basis of the technical scheme, the connection weight between the layers after correction is obtained according to the reverse transmission error from each data sample output layer to the hidden layer and the reverse transmission error from the hidden layer to the input layer, and the following formula is satisfied:
wherein E is total N is the number of detected samples, P is the known detection resultP is the data sample of the P-th known test result, and P e (1, 2,3,., P), K is the number of hidden layer neurons,target value of P-th neuron for hidden layer K-th layer,>hidden layer K-th layer P-th neuron output value, Δω jk For optimal connection weights of the kth neurons from the hidden layer to the jth layer of the output layer, deltav ij For the optimal connection weight of the j-th neuron from the input layer to the i-th layer of the hidden layer, eta is the learning rate,>and m is the number of neurons of the hidden layer.
On the basis of the technical scheme, the unstructured data comprise audio data, video data and picture data.
On the basis of the technical scheme, when the unstructured data are audio data, the unstructured data in the returned data are detected by adopting an artificial intelligent model, and the method specifically comprises the following steps of: firstly decoding the audio data, identifying the decoded audio frame, carrying out keyword matching search on the identified text information, and judging the harmful information of the audio data when the harmful information matched with the keyword reaches a preset threshold value during keyword matching.
On the basis of the technical scheme, when the unstructured data are video data, the unstructured data in the returned data are detected by adopting an artificial intelligent model, and the method specifically comprises the following steps of: firstly decoding video data, and performing frame extraction detection on the decoded video frames according to a specified rule, so as to judge whether the video data is harmful information or not.
In a second aspect, the invention further provides an artificial intelligence-based CDN harmful information blocking system, which comprises a log acquisition module, a log processing module and a monitoring task processing module;
the log acquisition module is used for acquiring request operation log data of a user in the CDN node and sending the request operation log data to the log processing module;
the log processing module is used for classifying, sequencing and de-duplication processing the request operation log data and sending the processed request operation log data to the monitoring task processing module;
the monitoring task processing module is used for judging the harmful information of the processed request operation log data, generating a blocking rule corresponding to the request operation log data which is judged to be the harmful information, and sending the blocking rule to all CDN nodes.
On the basis of the technical scheme, a plurality of monitoring task processing modules are arranged, and the log processing modules balance loads distributed to each monitoring task processing module through a load balancing algorithm.
Compared with the prior art, the invention has the advantages that:
according to the CDN harmful information blocking method based on artificial intelligence, detection of unstructured data such as audio, video and pictures is completed through an artificial intelligence model, and blocking and isolation of harmful information is completed according to detection results.
Drawings
Fig. 1 is a flow chart of a method for blocking CDN harmful information in an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
Referring to fig. 1, an embodiment of the present invention provides an artificial intelligence based CDN harmful information blocking method, which specifically includes the following steps:
acquiring request operation log data of users in all CDN nodes;
initiating a corresponding service request according to the request operation log data and acquiring returned data;
the method comprises the steps of obtaining a detection result of unstructured data in returned data by adopting an artificial intelligent model, and obtaining a detection result of the structured data in the returned data by keyword filtering and URL filtering, wherein the structured data is in a known format or data with a special flow to describe a protocol flow or an application operation flow, such as a media protocol request flow, an http request of standard resource query and the like, and the unstructured data is in an explicit flow or protocol format and can only be stored in the whole, such as pictures, audio and video data, various documents and the like.
And when the detection result is judged to be harmful information, generating a blocking rule of the request operation log data corresponding to the harmful information, and sending the blocking rule to all CDN nodes.
Because the artificial intelligent model can identify the yellow-involved and terrorist videos, the defect that the traditional CDN cannot detect video data is overcome, meanwhile, the traditional mode of manual auditing is avoided, the auditing flow of video resources is greatly simplified, the detection content is fully covered, and the detection precision is improved.
The artificial intelligent model adopts an artificial intelligent algorithm to finish detection of unstructured data, and when the artificial intelligent algorithm is adopted to finish detection of unstructured data, the artificial intelligent algorithm corrects the unstructured data through a plurality of known detection results, wherein when the artificial intelligent algorithm corrects the unstructured data, an artificial neural network BP (error back propagation) algorithm is adopted, and updating weights and threshold values are carried out through the unstructured data of a plurality of known detection results to finish correction of the algorithm, so that the final artificial intelligent algorithm can meet the accuracy of unstructured data detection.
The artificial intelligence algorithm comprises an input layer, an implicit layer and an output layer, and corrects the artificial intelligence algorithm through unstructured data of a plurality of known detection results, and specifically comprises the following steps:
initializing a connection weight from an input layer to an implicit layer unit and a connection weight from the implicit layer to an output layer;
sequentially inputting a plurality of data samples with known detection results to obtain the output of an implicit layer and an output layer of each data sample;
obtaining a reverse transmission error from each data sample output layer to the hidden layer and a reverse transmission error from the hidden layer to the input layer according to the output of each data sample hidden layer and the output layer;
and obtaining the connection weight between each layer after correction and the threshold value of the output layer according to the feedback error from each data sample output layer to the hidden layer and the feedback error from the hidden layer to the input layer. The threshold value is an empirical value, and is determined according to a scene, the higher the matching degree is, the larger the matching degree is, and the value range of the threshold value is [ 0,1 ].
Further, according to the output of each data sample hidden layer and output layer, the feedback error from each data sample output layer to the hidden layer and the feedback error from the hidden layer to the input layer are obtained, and the following formulas are satisfied:
δ k ＝(d k -0 k )f′(net k )
in delta k D, for outputting the feedback error from the layer to the hidden layer k To be the hidden layer K layer neuron target value (sample value), 0 k For the output layer, the Kth neuron outputs a value, f' (net k ) Is 0 to k And d k Deviation variance value, delta j For implicit layer to input layer feedback error, f' (net j ) For inputting the target value of the j-th layer neuron and the output value of the j-th layer neuron, omega jk The connection weight of the kth neuron from the hidden layer to the jth layer of the output layer.
Obtaining the connection weight between each layer after correction according to the reverse transmission error from each data sample output layer to the hidden layer and the reverse transmission error from the hidden layer to the input layer, and satisfying the following formula:
wherein E is total For the overall error, N is the number of detection samples, P is the number of data samples for a known detection result, P is the data sample for the P-th known detection result, and P e (1, 2, 3.., P), K is the number of layers of the hidden layer neurons,target value of P-th neuron for hidden layer K-th layer,>hidden layer K-th layer P-th neuron output value, Δω jk For optimal connection weights of the kth neurons from the hidden layer to the jth layer of the output layer, deltav ij For inputting the optimal connection weight of the j-th neuron from the layer to the i-th neuron of the hidden layer, further, delta omega jk And Deltav ij Respectively connecting weights from an implicit layer to an output layer and connecting weights from an input layer to the implicit layer, wherein eta is a learning rate, and further eta takes 0.5 #>And m is the number of neurons of the hidden layer.
Unstructured data includes audio data, video data, and picture data; when the unstructured data are audio data, firstly decoding the audio data, carrying out real-time speech translation recognition on the decoded audio frame, carrying out keyword matching search on the recognized text information, and judging the harmful information of the audio data when the harmful information matched with the keywords reaches a preset threshold value when the keywords are matched;
when the unstructured data are video data, the unstructured data in the returned data are detected by adopting an artificial intelligent model, and the method specifically comprises the following steps of: firstly decoding video data, performing frame extraction detection on the decoded video frames according to a specified rule, and further judging whether the video data is harmful information or not; when the detection of unstructured data is completed through an artificial intelligence algorithm, audio data, video data and picture data are respectively input into the corresponding corrected artificial intelligence algorithm, and according to the finally obtained result, the result is compared with the corresponding threshold value, and when the result obtained through the artificial intelligence algorithm reaches the threshold value, harmful information is judged.
When the unstructured data are picture data, the picture data are preprocessed in a mode of training picture preprocessing, the preprocessed picture is input into an optimal model formed by training, the model outputs obtained categories and category probabilities, and the categories with lower category probabilities are filtered through a threshold value to obtain a final detection result.
The embodiment of the invention also provides an artificial intelligence-based CDN harmful information blocking system, which comprises a log acquisition module, a log processing module and a monitoring task processing module;
the log acquisition module is used for acquiring request operation log data of a user in the CDN node and sending the request operation log data to the log processing module;
the log processing module is used for classifying, sequencing and de-duplication processing the request operation log data and sending the processed request operation log data to the monitoring task processing module;
the monitoring task processing module is used for judging the harmful information of the processed request operation log data, generating a blocking rule corresponding to the request operation log data judged to be the harmful information, and sending the blocking rule to all CDN nodes.
Through the one-time detection of the blocking system, the full-network blocking can be achieved, the detection result can be shared rapidly, conveniently and timely, meanwhile, the source can be traced back timely and accurately according to the tracking of the detection result source, and an accurate and powerful data basis is provided for the later-period resource integer and integration.
The monitoring task processing modules are provided with a plurality of log processing modules, so that the load distributed to each monitoring task processing module is balanced through a load balancing algorithm, and the reaction speed and the overall performance of the blocking system can be improved.
The monitoring task processing module comprises an audio monitoring sub-module, a video monitoring sub-module, a picture monitoring sub-module and a structured data monitoring sub-module, wherein the audio monitoring sub-module, the video monitoring sub-module and the picture monitoring sub-module respectively detect audio, video and pictures by adopting an artificial intelligent algorithm, and the structured data monitoring sub-module detects characters by means of keyword filtering and URL filtering.
The blocking system further includes a memory for storing the request operation log data, the data returned corresponding to the request operation log data, and the blocking rule.
When the log processing module classifies the request operation log data, an artificial intelligent algorithm is adopted for classification, and correction of the artificial intelligent algorithm is completed through the request operation log data of a known class, so that the accuracy of the log processing module in classifying the request operation log data is ensured.
The invention is not limited to the embodiments described above, but a number of modifications and adaptations can be made by a person skilled in the art without departing from the principle of the invention, which modifications and adaptations are also considered to be within the scope of the invention. What is not described in detail in this specification is prior art known to those skilled in the art.
Claims (5)
1. The CDN harmful information blocking system based on the artificial intelligence is characterized by adopting a CDN harmful information blocking method based on the artificial intelligence, and the method specifically comprises the following steps:
acquiring request operation log data of users in all CDN nodes;
initiating a corresponding service request according to the request operation log data and acquiring returned data;
the method comprises the steps of obtaining a detection result of unstructured data in returned data by adopting an artificial intelligent model, and obtaining the detection result of the structured data in the returned data by keyword filtering and URL filtering;
when the detection result judges that the information is harmful, generating a blocking rule of the request operation log data corresponding to the harmful information, and sending the blocking rule to all CDN nodes;
the unstructured data are audio data and video data;
when the unstructured data are audio data, the unstructured data in the returned data are detected by adopting an artificial intelligent model, and the method specifically comprises the following steps of: firstly decoding audio data, identifying the decoded audio frame, carrying out keyword matching search on the identified text information, and judging the audio data harmful information when the harmful information matched with the keyword reaches a preset threshold value during keyword matching;
when the unstructured data are video data, the unstructured data in the returned data are detected by adopting an artificial intelligent model, and the method specifically comprises the following steps of: firstly decoding video data, performing frame extraction detection on the decoded video frames according to a specified rule, and further judging whether the video data is harmful information or not;
the system comprises a log acquisition module, a log processing module and a monitoring task processing module;
the log acquisition module is used for acquiring request operation log data of a user in the CDN node and sending the request operation log data to the log processing module;
the log processing module is used for classifying, sequencing and de-duplication processing the request operation log data and sending the processed request operation log data to the monitoring task processing module;
the monitoring task processing module is used for judging the harmful information of the processed request operation log data, generating a blocking rule corresponding to the request operation log data which is judged to be the harmful information, and sending the blocking rule to all CDN nodes;
the plurality of monitoring task processing modules are arranged, and the log processing modules balance loads distributed to each monitoring task processing module through a load balancing algorithm;
the blocking system also comprises a memory, wherein the memory is used for storing request operation log data, returned data corresponding to the request operation log data and blocking rules;
when the log processing module classifies the request operation log data, an artificial intelligent algorithm is adopted for classification, and correction of the artificial intelligent algorithm is completed through the request operation log data with known types.
2. The CDN harmful information blocking system based on artificial intelligence of claim 1, wherein: the artificial intelligent model adopts an artificial intelligent algorithm to finish detection of unstructured data, and when the artificial intelligent algorithm is adopted to finish detection of unstructured data, the artificial intelligent algorithm corrects the unstructured data through a plurality of known detection results.
3. The CDN harmful information blocking system based on artificial intelligence of claim 2, wherein the artificial intelligence algorithm includes an input layer, an hidden layer, and an output layer, and the artificial intelligence algorithm corrects by unstructured data of a number of known detection results, comprising the steps of:
initializing a connection weight from an input layer to an implicit layer unit and a connection weight from the implicit layer to an output layer;
sequentially inputting a plurality of data samples with known detection results to obtain the output of an implicit layer and an output layer of each data sample;
obtaining a reverse transmission error from each data sample output layer to the hidden layer and a reverse transmission error from the hidden layer to the input layer according to the output of each data sample hidden layer and the output layer;
and determining a threshold value of the output layer, and obtaining the connection weight between the layers after correction according to the reverse transmission error from the output layer to the hidden layer and the reverse transmission error from the hidden layer to the input layer of each data sample.
4. The CDN harmful information blocking system based on artificial intelligence of claim 3, wherein a feedback error from each data sample output layer to an hidden layer and a feedback error from the hidden layer to an input layer are obtained from outputs of each data sample hidden layer and output layer, satisfying the following formula:
in (1) the->For outputting layer-to-hidden layer counter error, < ->For the hidden layer (k) th layer neuron target value, < ->Output value for the kth neuron of the output layer, < ->Is->And->Deviation variance value, & gt>For implicit layer to input layer counter error, < >>For inputting the jth layer neuron target value and the jth hidden layer neuron output value, +.>For the connection weights of the kth neurons of the hidden layer to the jth layer of the output layer,Kis the number of layers of hidden layer neurons.
5. The CDN harmful information blocking system of claim 4 wherein the corrected connection weights between layers are obtained from each of the data sample output layer to hidden layer feedback errors and hidden layer to input layer feedback errors, satisfying the following formula:
in (1) the->As a result of the overall error,Nto detect the number of samples，PFor the number of data samples for which the detection result is known,pis the firstpData samples of known test results, and，Kfor the number of layers of hidden layer neurons, +.>Target value of p-th neuron for the kth layer of the hidden layer,>hidden layer kth layer p-th neuron output value +.>Optimal connection weights for the kth neurons from the hidden layer to the jth layer of the output layer,/->For inputting the optimal connection weight of the j-th neuron of the layer to the i-th neuron of the hidden layer, is>For learning rate->For the number of neurons of the output layer,mfor the number of neurons in the hidden layer,ν ij the connection weight of the j-th neuron from the input layer to the i-th layer of the hidden layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911414631.7A CN111125539B (en) | 2019-12-31 | 2019-12-31 | CDN harmful information blocking method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911414631.7A CN111125539B (en) | 2019-12-31 | 2019-12-31 | CDN harmful information blocking method and system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111125539A CN111125539A (en) | 2020-05-08 |
| CN111125539B true CN111125539B (en) | 2024-02-02 |
Family
ID=70506641
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911414631.7A Active CN111125539B (en) | 2019-12-31 | 2019-12-31 | CDN harmful information blocking method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111125539B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2021270052A1 (en) * | 2020-05-09 | 2023-02-02 | Gt Systems Pty Ltd | Media distribution and management system and apparatus |
| CN113449782B (en) * | 2021-06-18 | 2022-05-24 | 中电积至(海南)信息技术有限公司 | CDN (content delivery network) hosting node detection method based on graph semi-supervised classification |
Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012152771A2 (en) * | 2011-05-12 | 2012-11-15 | Telefonica, S.A | Content server of a service provider's cdn |
| CN103634692A (en) * | 2013-12-06 | 2014-03-12 | 南京理工大学连云港研究院 | CDN (content distribution network) and P2P (peer-to-peer) based hybrid stream media video-on-demand system |
| CN105844238A (en) * | 2016-03-23 | 2016-08-10 | 乐视云计算有限公司 | Method and system for discriminating videos |
| WO2016164403A1 (en) * | 2015-04-10 | 2016-10-13 | Level 3 Communications, Llc | Systems and methods for generating network threat intelligence |
| CN106210057A (en) * | 2016-07-13 | 2016-12-07 | 成都知道创宇信息技术有限公司 | A kind of cloud security means of defence based on CDN |
| CN106778590A (en) * | 2016-12-09 | 2017-05-31 | 厦门大学 | It is a kind of that video detecting method is feared based on convolutional neural networks model cruelly |
| CN107211035A (en) * | 2014-12-05 | 2017-09-26 | 瑞典爱立信有限公司 | Method and network node for monitoring service in content delivery network |
| CN107992578A (en) * | 2017-12-06 | 2018-05-04 | 任明和 | The database automatic testing method in objectionable video source |
| CN108229535A (en) * | 2017-12-01 | 2018-06-29 | 百度在线网络技术（北京）有限公司 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
| CN108419091A (en) * | 2018-03-02 | 2018-08-17 | 北京未来媒体科技股份有限公司 | A kind of verifying video content method and device based on machine learning |
| CN108509649A (en) * | 2018-04-13 | 2018-09-07 | 郑俊杰 | A kind of violated Video content retrieval system |
| CN109376634A (en) * | 2018-10-15 | 2019-02-22 | 北京航天控制仪器研究所 | A kind of Bus driver unlawful practice detection system neural network based |
| CN109446461A (en) * | 2018-10-29 | 2019-03-08 | 成都思维世纪科技有限责任公司 | A kind of method of CDN and CACHE caching flame content auditing |
| CN109508402A (en) * | 2018-11-15 | 2019-03-22 | 上海指旺信息科技有限公司 | Violation term detection method and device |
| CN109754306A (en) * | 2018-11-13 | 2019-05-14 | 北京码牛科技有限公司 | Information processing method and device, electronic equipment and computer-readable medium |
| KR101993158B1 (en) * | 2019-01-03 | 2019-06-27 | 넷마블 주식회사 | Method and apparatus for selecting contents delivary network |
| CN110210542A (en) * | 2019-05-24 | 2019-09-06 | 厦门美柚信息科技有限公司 | Picture character identification model training method, device and character identification system |
| CN110532866A (en) * | 2019-07-22 | 2019-12-03 | 平安科技（深圳）有限公司 | Video data detection method, device, computer equipment and storage medium |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160344751A1 (en) * | 2015-05-19 | 2016-11-24 | Fastly, Inc. | Customized record handling in a content delivery network |
| US10237286B2 (en) * | 2016-01-29 | 2019-03-19 | Zscaler, Inc. | Content delivery network protection from malware and data leakage |
| US20170331913A1 (en) * | 2016-05-11 | 2017-11-16 | Fastly, Inc. | Content based content delivery network purging |
-
2019
- 2019-12-31 CN CN201911414631.7A patent/CN111125539B/en active Active
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012152771A2 (en) * | 2011-05-12 | 2012-11-15 | Telefonica, S.A | Content server of a service provider's cdn |
| CN103634692A (en) * | 2013-12-06 | 2014-03-12 | 南京理工大学连云港研究院 | CDN (content distribution network) and P2P (peer-to-peer) based hybrid stream media video-on-demand system |
| CN107211035A (en) * | 2014-12-05 | 2017-09-26 | 瑞典爱立信有限公司 | Method and network node for monitoring service in content delivery network |
| WO2016164403A1 (en) * | 2015-04-10 | 2016-10-13 | Level 3 Communications, Llc | Systems and methods for generating network threat intelligence |
| CN105844238A (en) * | 2016-03-23 | 2016-08-10 | 乐视云计算有限公司 | Method and system for discriminating videos |
| CN106210057A (en) * | 2016-07-13 | 2016-12-07 | 成都知道创宇信息技术有限公司 | A kind of cloud security means of defence based on CDN |
| CN106778590A (en) * | 2016-12-09 | 2017-05-31 | 厦门大学 | It is a kind of that video detecting method is feared based on convolutional neural networks model cruelly |
| CN108229535A (en) * | 2017-12-01 | 2018-06-29 | 百度在线网络技术（北京）有限公司 | Relate to yellow image audit method, apparatus, computer equipment and storage medium |
| CN107992578A (en) * | 2017-12-06 | 2018-05-04 | 任明和 | The database automatic testing method in objectionable video source |
| CN108419091A (en) * | 2018-03-02 | 2018-08-17 | 北京未来媒体科技股份有限公司 | A kind of verifying video content method and device based on machine learning |
| CN108509649A (en) * | 2018-04-13 | 2018-09-07 | 郑俊杰 | A kind of violated Video content retrieval system |
| CN109376634A (en) * | 2018-10-15 | 2019-02-22 | 北京航天控制仪器研究所 | A kind of Bus driver unlawful practice detection system neural network based |
| CN109446461A (en) * | 2018-10-29 | 2019-03-08 | 成都思维世纪科技有限责任公司 | A kind of method of CDN and CACHE caching flame content auditing |
| CN109754306A (en) * | 2018-11-13 | 2019-05-14 | 北京码牛科技有限公司 | Information processing method and device, electronic equipment and computer-readable medium |
| CN109508402A (en) * | 2018-11-15 | 2019-03-22 | 上海指旺信息科技有限公司 | Violation term detection method and device |
| KR101993158B1 (en) * | 2019-01-03 | 2019-06-27 | 넷마블 주식회사 | Method and apparatus for selecting contents delivary network |
| CN110210542A (en) * | 2019-05-24 | 2019-09-06 | 厦门美柚信息科技有限公司 | Picture character identification model training method, device and character identification system |
| CN110532866A (en) * | 2019-07-22 | 2019-12-03 | 平安科技（深圳）有限公司 | Video data detection method, device, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 袁冰清等.BP神经网络基本原理.《数字通信世界》.2018,(第undefined期),28-29. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111125539A (en) | 2020-05-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220179620A1 (en) | System and method for enriching datasets while learning | |
| CN113255755B (en) | Multi-modal emotion classification method based on heterogeneous fusion network | |
| US11210595B2 (en) | System and method for selective use of examples | |
| US11823658B2 (en) | Trial-based calibration for audio-based identification, recognition, and detection system | |
| JP2022536189A (en) | Method, Apparatus, Equipment and Storage Medium for Recognizing Voiceprint of Original Speech | |
| CN105931633A (en) | Speech recognition method and system | |
| CN111125539B (en) | CDN harmful information blocking method and system based on artificial intelligence | |
| CN114743020A (en) | Food identification method combining tag semantic embedding and attention fusion | |
| WO2017183242A1 (en) | Information processing device and information processing method | |
| US20150161107A1 (en) | Discriminating Between Natural Language and Keyword Language Items | |
| JP2022508091A (en) | Dynamic reconstruction training computer architecture | |
| Shang et al. | Fedic: Federated learning on non-iid and long-tailed data via calibrated distillation | |
| WO2021092631A2 (en) | Weakly-supervised text-based video moment retrieval | |
| CN117079299B (en) | Data processing method, device, electronic equipment and storage medium | |
| WO2021092632A2 (en) | Weakly-supervised text-based video moment retrieval via cross attention modeling | |
| Marshall et al. | A neural network approach for truth discovery in social sensing | |
| CN115393968A (en) | Audio-visual event positioning method fusing self-supervision multi-mode features | |
| US11899765B2 (en) | Dual-factor identification system and method with adaptive enrollment | |
| CN113935398A (en) | Network traffic classification method and system based on small sample learning in Internet of things environment | |
| Drishya et al. | Cyberbully image and text detection using convolutional neural networks | |
| CN104572820B (en) | The generation method and device of model, importance acquisition methods and device | |
| CN116977701A (en) | Video classification model training method, video classification method and device | |
| Llopis et al. | Matching user queries in natural language with Cyber-Physical Systems using deep learning through a Transformer approach | |
| CN115643153A (en) | Alarm correlation analysis method based on graph neural network | |
| CN112069806B (en) | Resume screening method and device, electronic equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |