CN112528317A - Information processing method, device and equipment based on block chain - Google Patents
Information processing method, device and equipment based on block chain Download PDFInfo
- Publication number
- CN112528317A CN112528317A CN202011244859.9A CN202011244859A CN112528317A CN 112528317 A CN112528317 A CN 112528317A CN 202011244859 A CN202011244859 A CN 202011244859A CN 112528317 A CN112528317 A CN 112528317A
- Authority
- CN
- China
- Prior art keywords
- participant
- participants
- block chain
- parameters
- artificial intelligence
- 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
Links
- 230000010365 information processing Effects 0.000 title claims abstract description 28
- 238000003672 processing method Methods 0.000 title claims abstract description 23
- 238000012549 training Methods 0.000 claims abstract description 63
- 238000013473 artificial intelligence Methods 0.000 claims abstract description 58
- 230000002159 abnormal effect Effects 0.000 claims abstract description 46
- 230000006399 behavior Effects 0.000 claims abstract description 44
- 238000012544 monitoring process Methods 0.000 claims abstract description 13
- 238000012795 verification Methods 0.000 claims description 29
- 238000000034 method Methods 0.000 claims description 28
- 238000004891 communication Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000009825 accumulation Methods 0.000 claims description 3
- 230000008260 defense mechanism Effects 0.000 abstract description 6
- 238000012550 audit Methods 0.000 abstract description 3
- 238000001514 detection method Methods 0.000 abstract description 3
- 230000002776 aggregation Effects 0.000 description 8
- 238000004220 aggregation Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 230000015572 biosynthetic process Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 238000003786 synthesis reaction Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
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
- G06F21/60—Protecting data
- G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
- G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The invention discloses an information processing method, device and equipment based on a block chain, which are used for monitoring operation behavior parameters of a participant who utilizes the block chain to carry out artificial intelligence model training in each preset time period, wherein blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters. Therefore, if the operation behavior parameter difference between a certain participant and other participants is large, the participant is considered to be abnormal in operation, corresponding operation is adopted, the situation that the parameters obtained by abnormal operation behaviors damage the global model on the block chain is effectively avoided, attack of the participant on the block chain is effectively defended, meanwhile, detection of the participant is recorded on the block chain, all the participants can audit a defense mechanism, the rationality of the block chain defense mechanism is guaranteed, and the safety and reliability of artificial intelligence model training of the block chain are remarkably improved.
Description
Technical Field
The present invention relates to the field of block chain technologies, and in particular, to a method, an apparatus, and a device for processing information based on a block chain.
Background
As is well known, artificial intelligence is also continuously developed at a rapid and rapid pace under the promotion of information calculation and network communication, but the artificial intelligence is obtained by continuously calculating and learning an artificial intelligence model based on a large amount of data, and the quality and scale of training data of the artificial intelligence model directly determine the excellent degree of the artificial intelligence.
Blockchains are an information technology emerging in recent years. In essence, the blockchain can be regarded as a shared database, and the data or information stored therein has the characteristics of being unforgeable, having no trace in the whole process, having traceability, having public transparency, having collective maintenance, and the like. Based on the characteristics, the block chain technology lays a solid 'trust' foundation, creates a reliable 'cooperation' mechanism and has wide application prospect.
As blockchain technology continues to mature, researchers in the neighborhood of artificial intelligence have also attempted to train out more powerful artificial intelligence through data generated by the blockchain market. Research on a block chain-based artificial intelligence model training scenario BDML (block-based Distributed Machine Learning), and specifically, in order to solve a specific artificial intelligence problem, a block chain technology is applied to cooperatively train a model under the condition of protecting data of each participant from being leaked.
However, since the local private data of each participant is not shared, there may be a countermeasure attack on the BDML network by a partially malicious participant through a data attack or model attack, for example, mixing a wrong example in the local data set, or modifying the local model parameters so that the local model is biased, and if the local model of the malicious participant is aggregated by other participants, the global model on the block chain may be damaged, so that the global model is biased.
Disclosure of Invention
In order to solve the above problems, embodiments of the present invention creatively provide an information processing method, apparatus, and device based on a block chain.
According to a first aspect of the present invention, there is provided a block chain-based information processing method, the method including: monitoring operation behavior parameters of a participant who utilizes a block chain to perform artificial intelligence model training in each preset time period, wherein blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters.
According to an embodiment of the invention, the method further comprises: for one participant, accumulating the abnormal times of the participant determined as abnormal operation; when the abnormal times reach a set abnormal threshold value, confirming that the operation of the participant is abnormal; moving the participant out of the blockchain.
According to an embodiment of the invention, the operational behavior parameter comprises at least one of: the participator is used as an artificial intelligence model trained by the verifier to other participators, and the ratio of the number of times of verification passing of other participators to the number of times of verification failure of other participators is judged; the verification passing rate of the artificial intelligence model trained by the participant as a generator and verified by other participants; the parameters trained and sent by the participants are different from the training parameters of the model parameters on the first block generated on the block chain after the parameters are sent; an aggregate probability that the parameters issued by the participants are aggregated by other participants.
According to an embodiment of the present invention, the determining, according to the operation behavior parameter, whether the participant operates abnormally within a corresponding preset time period includes: when the operation behavior parameters meet at least one of the following conditions, judging that the operation of the participant is abnormal within a corresponding preset time period: the difference between the voted passage ratio of the participant and the mean of the voted passage ratios of the other participants is greater than a first set threshold; the difference between the verified passing rate of the participant and the mean of the verified passing rates of the other participants is greater than a second set threshold; the difference between the training parameter difference of the participant and the mean of the training parameter differences of the other participants is greater than a third set threshold; the difference between the aggregate probability of the participant and the mean of the aggregate probabilities of the other participants is greater than a fourth set threshold.
According to an embodiment of the present invention, the preset time period is a time period between generation times of two adjacent blocks in the block chain.
According to a second aspect of the embodiments of the present invention, there is also provided an information processing apparatus based on a block chain, the apparatus including: the monitoring module is used for monitoring the operation behavior parameters of a participant who utilizes a block chain to perform artificial intelligence model training in each preset time period, and blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and the abnormity judging module is used for judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters.
According to an embodiment of the invention, the apparatus further comprises: the accumulation module is used for accumulating the abnormal times of the participants which are judged to be abnormal in operation aiming at one participant; the abnormity confirming module is used for confirming that the operation of the participant is abnormal when the abnormity frequency reaches a set abnormity threshold value; and the exception handling module is used for moving the participant out of the block chain.
According to an embodiment of the invention, the operational behavior parameter comprises at least one of: the participator is used as an artificial intelligence model trained by the verifier to other participators, and the ratio of the number of times of verification passing of other participators to the number of times of verification failure of other participators is judged; the verification passing rate of the artificial intelligence model trained by the participant as a generator and verified by other participants; the parameters trained and sent by the participants are different from the training parameters of the model parameters on the first block generated on the block chain after the parameters are sent; an aggregate probability that the parameters issued by the participants are aggregated by other participants.
According to an embodiment of the present invention, the abnormality determination module includes: the parameter judgment submodule is used for judging that the operation of the participant is abnormal within a corresponding preset time period when the operation behavior parameter meets at least one of the following conditions: the difference between the voted passage ratio of the participant and the mean of the voted passage ratios of the other participants is greater than a first set threshold; the difference between the verified passing rate of the participant and the mean of the verified passing rates of the other participants is greater than a second set threshold; the difference between the training parameter difference of the participant and the mean of the training parameter differences of the other participants is greater than a third set threshold; the difference between the aggregate probability of the participant and the mean of the aggregate probabilities of the other participants is greater than a fourth set threshold.
According to a third aspect of the present invention, there is also provided a computer-readable storage medium comprising a set of computer-executable instructions, which when executed, perform any of the above-mentioned methods for processing blockchain-based information.
According to a fourth aspect of the present invention, there is also provided an apparatus comprising at least one processor, and at least one memory connected to the processor, a bus; the processor and the memory complete mutual communication through the bus; the processor is used for calling the program instructions in the memory so as to execute the information processing method based on the block chain.
The embodiment of the invention discloses an information processing method, device and equipment based on a block chain, which are used for monitoring the operation behavior parameters of a participant who utilizes the block chain to carry out artificial intelligence model training in each preset time period, wherein the blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters. Therefore, if the operation behavior parameter difference between a certain participant and other participants is large, the participant is considered to be abnormal in operation, corresponding operation is adopted, the situation that the parameters obtained by abnormal operation behaviors damage the global model on the block chain is effectively avoided, attack of the participant on the block chain is effectively defended, meanwhile, detection of the participant is recorded on the block chain, all the participants can audit a defense mechanism, the rationality of the block chain defense mechanism is guaranteed, and the safety and reliability of artificial intelligence model training of the block chain are remarkably improved.
It is to be understood that the teachings of the present invention need not achieve all of the above-described benefits, but rather that specific embodiments may achieve specific technical results, and that other embodiments of the present invention may achieve benefits not mentioned above.
Drawings
The above and other objects, features and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description read in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which:
in the drawings, the same or corresponding reference numerals indicate the same or corresponding parts.
Fig. 1 is a schematic flow chart illustrating an implementation of an information processing method based on a block chain according to an embodiment of the present invention;
FIG. 2 is a first schematic flow chart illustrating an implementation example of an application of the block chain-based information processing method according to the embodiment of the present invention;
fig. 3 is a schematic diagram illustrating an implementation flow of an application example of the information processing method based on the block chain according to the embodiment of the present invention;
fig. 4 is a schematic diagram showing a third implementation flow of an application example of the information processing method based on the block chain according to the embodiment of the present invention;
FIG. 5 is a block chain-based information processing apparatus according to an embodiment of the present invention;
fig. 6 is a schematic diagram showing a composition structure of the apparatus according to the embodiment of the present invention.
Detailed Description
The principles and spirit of the present invention will be described with reference to a number of exemplary embodiments. It is understood that these embodiments are given only to enable those skilled in the art to better understand and to implement the present invention, and do not limit the scope of the present invention in any way. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
The technical solution of the present invention is further elaborated below with reference to the drawings and the specific embodiments.
Fig. 1 is a schematic diagram illustrating an implementation flow of an information processing method based on a block chain according to an embodiment of the present invention.
Referring to fig. 1, an information processing method based on a block chain according to an embodiment of the present invention at least includes the following operation flows: operation 101, monitoring operation behavior parameters of a participant performing artificial intelligence model training by using a block chain in each preset time period, wherein blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and operation 102, judging whether the operation of the participant is abnormal within a corresponding preset time period according to the operation behavior parameters.
In operation 101, the operational behavior parameters of the participant performing artificial intelligence model training by using the blockchain are monitored in each preset time period, and the blocks on the blockchain are generated when the artificial intelligence model trained by the participant is verified by other participants.
In an embodiment of the present invention, the preset time period is a time period between generation times of two adjacent blocks in the block chain.
When the artificial intelligence model training is carried out based on the block chain, each participant can be used as a verifier to verify and vote on the artificial intelligence model trained by other participants. Each participant can also be used as a model training party to perform model training, and when the artificial intelligence model trained by other verification parties passes, a new block is generated on the block chain. For example: participants were trained separately on their respective training data sets using random gradient descent (SGD). The data set of the user is divided into a plurality of batchs according to the preset Batch Size. And training by using all batchs in each round according to a preset training round number Epoch to finally obtain a parameter vector w. And the participator tests the local test data set of the participator to find that the local parameters are superior to the parameters of the latest block on the block chain, and then broadcasts the model parameters to all other participators. And the other participants verify the model trained by the participant by using the corresponding local data respectively, and give a voting result, and when the verification is passed, a new block is generated on the block chain. The time duration of the time period between the generation times of two adjacent blocks in the block chain may be the same or different, and usually, the time duration of the time period between the generation times of two adjacent blocks may have a certain difference.
In one embodiment of the invention, the operational behavior parameter comprises at least one of: the participator is used as an artificial intelligence model trained by the verifier to other participators, and the vote passing proportion of the times of passing verification of other participators and the times of failing verification of other participators is judged; the verification passing rate of the artificial intelligence model trained by the participant as the generator to be verified by other participants; the parameters trained and sent by the participants are different from the training parameters of the model parameters on the first block generated on the block chain after the parameters are sent; aggregate probability that parameters issued by a participant are aggregated by other participants.
For example, when the artificial intelligence model is trained based on the blockchain, each participant can be used as a verifier to verify and vote on the artificial intelligence model trained by other participants. For example: the participant P1 votes on the parameter W2 of the model M2 trained by the participant P2, and then the participant P1 determines that the parameter W2 of the model M2 trained by the participant P2 passes the verification, and on the contrary, the participant P1 determines that the parameter W2 of the model M2 trained by the participant P2 fails the verification. Therefore, the artificial intelligence model trained by the participant P1 as the verifier to other participants can be monitored in each preset time period, and the ratio of the number of verification passes of other participants to the number of verification failures of other participants is determined.
Similarly, any participant performing artificial intelligence model training based on the blockchain can also perform model training as a generator, and the trained artificial intelligence model needs to be verified by other participants, and a new block is generated in the blockchain after the verification is passed. In each preset time period, the participator can perform multiple rounds of artificial intelligence model training, and after each round of training is completed, the participator can receive the verification results of other participators on the model, for example: approval or disapproval of the ticket. The ratio of the number of participants who have cast the negative vote to all participants who have cast positive and negative votes is the verified passing rate. If the participant Pi performs multiple rounds of model training within a preset time period, the verified passing rate may be the mean of the verified passing rates of the multiple rounds of model training.
For the training parameter difference between the parameter trained and sent by the participant and the model parameter on the first block generated on the block chain after the parameter is sent, for example: and a new block is generated on the block chain at the time of T0, the model parameter on the new block is W2, the participant P3 obtains W2, model training is completed at the time of T1 based on W2, and the model parameter W3 is obtained, but the verification of other participants is not passed. And the model parameter W5 trained by the participant P5 at the later time T2 is verified by other participants to generate a new block, and no other block is generated between the times T0 and T2, so that the training parameter difference between the parameter trained and sent by the participant P2 and the model parameter on the first block generated on the block chain after the parameter is sent is the difference between W3 and W5 in the preset time period from the time T0 to T2.
For example, in the course of training the artificial intelligence model based on the blockchain, the model parameter WA1 trained by the participant a is not verified by other participants, and at the same time of training the model by the participant a, the participant M, N, L also performs model training, and accordingly the trained model parameters are WM1, WN1 and WL1 in sequence, but all of them are not verified by other participants, and the verification result of the participant a on the participant M, N, L is in turn approved, approved and disapproved. Participant a may aggregate the model parameters WM1, WN1, WL1 of participant M, N, L and perform model training again. Participant a may also select the model parameters WM1 trained by participant M and WN1 trained by participant N who participant a voted on for model training again. In practical application, the number of the participants is large, and the participant a can also adopt other aggregation rules to synthesize the model participation trained by other participants, and then perform model training again. Thus, the number of times that each participant is subjected to parameter synthesis by other participants within a preset time period can be recorded and accumulated, for example: the parameters of the participant M are integrated by the participant a, and the number of times that the participant M is integrated by other participants within the preset time period is increased by 1. The number of times that the participant has integrated the participation of other participants within a preset time period may also be recorded and accumulated, for example: and the participant A integrates the parameters of the participants M and N to carry out model training, so that the times of integrating other participants by the participant in a preset time period is increased by 1. The ratio of the number of times that each participant is subjected to parameter synthesis by other participants within a preset time period to the number of times that the participant synthesizes other participants is recorded as the aggregation probability that the parameters sent by the participants are aggregated by other participants, for example: the parameters of the party M are integrated by the parties, such as party a, 2 times during a preset time period TX-TY, and the parties, such as party a, perform model training for 8 times in total during the preset time period. The aggregate probability of the parameters issued by party M being aggregated by the other parties is 2/8 between the preset time periods TX-TY.
In operation 102, whether the participant operates abnormally within a corresponding preset time period is determined according to the operation behavior parameters.
In an embodiment of the present invention, determining whether a participant operates abnormally within a corresponding preset time period according to an operation behavior parameter includes: when the operation behavior parameters meet at least one of the following conditions, judging that the operation of the participant is abnormal within the corresponding preset time period: the difference value between the vote passing proportion of the participant and the mean value of the vote passing proportions of other participants is larger than a first set threshold value; the difference value between the verified passing rate of the participant and the mean value of the verified passing rates of other participants is larger than a second set threshold value; the difference value between the training parameter difference of the participant and the mean value of the training parameter differences of other participants is larger than a third set threshold value; the difference between the aggregation probability of the participant and the mean of the aggregation probabilities of the other participants is greater than a fourth set threshold.
It should be noted that the first set threshold, the second set threshold, the third set threshold, and the fourth set threshold may be determined values or values that vary with other parameters. The first set threshold, the second set threshold, the third set threshold and the fourth set threshold may be the same or different, or any two of them may be the same, which is not limited in the present invention.
In one embodiment of the invention, the abnormal times of the participator which is determined as abnormal operation are accumulated for one participator; when the abnormal times reach a set abnormal threshold value, confirming that the operation of the participant is abnormal; the participant is moved out of the blockchain.
The abnormal threshold may be set according to the number of blocks in the current blockchain, for example: setting the exception threshold to 1/4, which is the number of blocks in the current blockchain, may also be a fixed value, for example: 3.
for example, when 17 blocks are in the front blockchain, and the accumulated abnormal times of the participator a being determined as abnormal operation is 5, the participator a is determined to be abnormal operation, and the participator is removed from the blockchain.
In an embodiment of the present invention, the participant confirmed as the abnormal operation may be marked as an abnormal participant, and the participant may be prohibited from generating a new block.
In one embodiment of the present invention, the broadcast is also broadcast to all participants when it is determined that the participants are operating abnormally within some preset time period. Further, the participant identification confirmed as abnormal operation can be broadcasted to inform other participants of the trained model parameters and the submitted model verification of the participant.
FIG. 2 is a first schematic flow chart illustrating an implementation example of an application of the block chain-based information processing method according to the embodiment of the present invention; fig. 3 is a schematic diagram illustrating an implementation flow of an application example of the information processing method based on the block chain according to the embodiment of the present invention; fig. 4 is a schematic diagram showing an implementation flow of an application example of the information processing method based on the block chain according to the embodiment of the present invention.
Referring to fig. 2-4, in an application example of the information processing method based on a blockchain according to the embodiment of the present invention, any participant Pi may perform training of an artificial intelligence model based on the blockchain according to the following operation steps in fig. 2.
In operation 201, the latest parameter Wc on the blockchain is pulled.
In operation 202, a parameter aggregation policy p is preset.
At operation 203, the parameters Wi are trained using the self-contained training data.
At operation 204, testing is performed using the owned test data.
At operation 205, the accuracy is better than Wc.
At operation 206, the parameters Wi are packaged and broadcast to other participants.
At operation 207, the other participating parties vote.
In operation 209, the parameters in Wr are aggregated according to the aggregation policy p to obtain the aggregated Wi.
Likewise, any participant Pi can also verify the artificial intelligence model trained by other participants based on the blockchain according to the following operation steps in fig. 3.
While performing operation 206, the voting passage rate of the party Pi required to be monitored by operation 4031 within the preset time window is monitored.
In operation 207, the party to be monitored in monitoring operation 4032 issues a difference between the parameter and the new sector block parameter within a preset time window.
In operation 208, the probability that the parameters sent out by each participant within the preset time window, which are required to be counted in operation 4033, are aggregated is counted.
In operations 304-306, the parameter Pi that is required to be counted in the counting operation 4034 is counted to determine the ratio between positive and negative votes within a predetermined time window.
Operation 401, determine a preset time window, an abnormal threshold T and a threshold TV、TD、TP、TR. The preset time window is the preset time period in the above operations 101-102.
At operation 402, the operation behavior of the Pi participants within a preset time window is detected.
In operation 4031, a vote pass rate V is determined.
In operation 4032, the difference D between the issued parameter and the new block parameter.
In operation 4033, a probability P that the parameter is aggregated.
In operation 4034, a ratio R of positive to negative tickets is cast.
In operation 4041, the vote pass rate V differs from the mean by more than TV。
At operation 4042, the difference D and the mean of the issued parameter and the new block parameter exceed TD。
Operation 4043, a probability P that the parameters are aggregated differs from the mean by more than TP。
Operation 4044, the ratio R of positive to negative votes cast differs by more than T from the meanR。
The specific implementation process of operations 4031 to 4044 is performed once for one participating party, and the specific implementation flow is similar to the specific implementation process of operations 101 and 102 in the embodiment shown in fig. 1, and is not described again here.
At operation 405, the outlier of party Pi is incremented by 1.
At operation 406, party Pi exceeds threshold T.
At operation 407, the participant Pi is moved out of the blockchain network.
The details of the implementation processes of operations 101 to 102 in fig. 1 may be referred to for the content not shown in the application example of the embodiment of the present invention, and are not described herein again.
The embodiment of the invention discloses an information processing method, device and equipment based on a block chain, which are used for monitoring the operation behavior parameters of a participant who utilizes the block chain to carry out artificial intelligence model training in each preset time period, wherein blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters. Therefore, if the operation behavior parameter difference between a certain participant and other participants is large, the participant is considered to be abnormal in operation, corresponding operation is adopted, the situation that the parameters obtained by abnormal operation behaviors damage the global model on the block chain is effectively avoided, attack of the participant on the block chain is effectively defended, meanwhile, detection of the participant is recorded on the block chain, all the participants can audit a defense mechanism, the rationality of the block chain defense mechanism is guaranteed, and the safety and reliability of artificial intelligence model training of the block chain are remarkably improved.
Similarly, based on the above block chain-based information processing method, an embodiment of the present invention further provides a computer-readable storage medium, in which a program is stored, and when the program is executed by a processor, the processor is caused to perform at least the following operation steps: operation 101, monitoring operation behavior parameters of a participant performing artificial intelligence model training by using a block chain in each preset time period, wherein blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants; and operation 102, judging whether the operation of the participant is abnormal within a corresponding preset time period according to the operation behavior parameters.
Further, based on the above block chain based information processing method, an embodiment of the present invention further provides a block chain based information processing apparatus, as shown in fig. 5, where the apparatus 50 includes: the monitoring module 501 is configured to monitor an operation behavior parameter of a participant performing artificial intelligence model training by using a blockchain in each preset time period, where a block on the blockchain is generated when the artificial intelligence model trained by the participant is verified by other participants; and an anomaly determination module 502, configured to determine whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameter.
In one embodiment of the present invention, the apparatus 50 further comprises: the accumulation module is used for accumulating the abnormal times of the participants which are judged to be abnormal in operation aiming at one participant; the abnormity confirming module is used for confirming that the operation of the participant is abnormal when the abnormity frequency reaches a set abnormity threshold value; and the exception handling module is used for moving the participant out of the block chain.
In one embodiment of the invention, the operational behavior parameter comprises at least one of: the participator is used as an artificial intelligence model trained by the verifier to other participators, and the vote passing proportion of the times of passing verification of other participators and the times of failing verification of other participators is judged; the verification passing rate of the artificial intelligence model trained by the participant as the generator to be verified by other participants; the parameters trained and sent by the participants are different from the training parameters of the model parameters on the first block generated on the block chain after the parameters are sent; aggregate probability that parameters issued by a participant are aggregated by other participants.
In an embodiment of the present invention, the abnormality determining module 502 includes: the parameter judgment submodule is used for judging that the operation of the participant is abnormal in the corresponding preset time period when the operation behavior parameter meets at least one of the following conditions: the difference value between the vote passing proportion of the participant and the mean value of the vote passing proportions of other participants is larger than a first set threshold value; the difference value between the verified passing rate of the participant and the mean value of the verified passing rates of other participants is larger than a second set threshold value; the difference value between the training parameter difference of the participant and the mean value of the training parameter differences of other participants is larger than a third set threshold value; the difference between the aggregation probability of the participant and the mean of the aggregation probabilities of the other participants is greater than a fourth set threshold.
Still further, based on the above block chain-based information processing method, the present invention further provides an apparatus, and referring to fig. 6, the apparatus 60 includes at least one processor 601, and at least one memory 602 and a bus 603 connected to the processor 601; the processor 601 and the memory 602 complete communication with each other through the bus 603; the processor 601 is used for calling program instructions in the memory 602 to execute the above-mentioned information processing method based on the blockchain.
Here, it should be noted that: the above description of the embodiments of the method and the apparatus for processing information based on a block chain is similar to the description of the embodiments of the method shown in fig. 1 to 4, and has similar beneficial effects to the embodiments of the method shown in fig. 1 to 4, and therefore, the description is omitted here for brevity. For technical details that are not disclosed in the embodiments of the block chain information processing apparatus and device of the present invention, please refer to the description of the method embodiments shown in fig. 1 to 4 of the present invention for understanding, and therefore, for brevity, will not be described again.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described device embodiments are merely illustrative, for example, the division of a unit is only one logical function division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units; can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, all the functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: various media that can store program codes, such as a removable Memory device, a Read Only Memory (ROM), a magnetic disk, or an optical disk.
Alternatively, the integrated unit of the present invention may be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or a part contributing to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the methods of the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, a ROM, a magnetic or optical disk, or other various media that can store program code.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A method of block chain based information processing, the method comprising:
monitoring operation behavior parameters of a participant who utilizes a block chain to perform artificial intelligence model training in each preset time period, wherein blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants;
and judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters.
2. The method of claim 1, further comprising:
for one participant, accumulating the abnormal times of the participant determined as abnormal operation;
when the abnormal times reach a set abnormal threshold value, confirming that the operation of the participant is abnormal;
moving the participant out of the blockchain.
3. The method of claim 1, the operational behavior parameter comprising at least one of:
the participator is used as an artificial intelligence model trained by the verifier to other participators, and the ratio of the number of times of verification passing of other participators to the number of times of verification failure of other participators is judged;
the verification passing rate of the artificial intelligence model trained by the participant as a generator and verified by other participants;
the parameters trained and sent by the participants are different from the training parameters of the model parameters on the first block generated on the block chain after the parameters are sent;
an aggregate probability that the parameters issued by the participants are aggregated by other participants.
4. The method of claim 3, wherein the determining whether the participant operates abnormally within a corresponding preset time period according to the operational behavior parameters comprises: when the operation behavior parameters meet at least one of the following conditions, judging that the operation of the participant is abnormal within a corresponding preset time period:
the difference between the voted passage ratio of the participant and the mean of the voted passage ratios of the other participants is greater than a first set threshold;
the difference between the verified passing rate of the participant and the mean of the verified passing rates of the other participants is greater than a second set threshold;
the difference between the training parameter difference of the participant and the mean of the training parameter differences of the other participants is greater than a third set threshold;
the difference between the aggregate probability of the participant and the mean of the aggregate probabilities of the other participants is greater than a fourth set threshold.
5. The method according to any of claims 1-4, wherein the preset time period is a time period between generation times of two adjacent tiles on the tile chain.
6. An apparatus for processing block chain-based information, the apparatus comprising:
the monitoring module is used for monitoring the operation behavior parameters of a participant who utilizes a block chain to perform artificial intelligence model training in each preset time period, and blocks on the block chain are generated when the artificial intelligence model trained by the participant is verified by other participants;
and the abnormity judging module is used for judging whether the participant operates abnormally within a corresponding preset time period according to the operation behavior parameters.
7. The apparatus of claim 6, the apparatus further comprising:
the accumulation module is used for accumulating the abnormal times of the participants which are judged to be abnormal in operation aiming at one participant;
the abnormity confirming module is used for confirming that the operation of the participant is abnormal when the abnormity frequency reaches a set abnormity threshold value;
and the exception handling module is used for moving the participant out of the block chain.
8. The apparatus of claim 6, the operational behavior parameter comprising at least one of:
the participator is used as an artificial intelligence model trained by the verifier to other participators, and the ratio of the number of times of verification passing of other participators to the number of times of verification failure of other participators is judged;
the verification passing rate of the artificial intelligence model trained by the participant as a generator and verified by other participants;
the parameters trained and sent by the participants are different from the training parameters of the model parameters on the first block generated on the block chain after the parameters are sent;
an aggregate probability that the parameters issued by the participants are aggregated by other participants.
9. The apparatus of claim 8, the anomaly determination module comprising:
the parameter judgment submodule is used for judging that the operation of the participant is abnormal within a corresponding preset time period when the operation behavior parameter meets at least one of the following conditions:
the difference between the voted passage ratio of the participant and the mean of the voted passage ratios of the other participants is greater than a first set threshold;
the difference between the verified passing rate of the participant and the mean of the verified passing rates of the other participants is greater than a second set threshold;
the difference between the training parameter difference of the participant and the mean of the training parameter differences of the other participants is greater than a third set threshold;
the difference between the aggregate probability of the participant and the mean of the aggregate probabilities of the other participants is greater than a fourth set threshold.
10. A device comprising at least one processor, and at least one memory, bus connected with the processor; the processor and the memory complete mutual communication through the bus; the processor is used for calling the program instructions in the memory to execute the block chain-based information processing method of any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011244859.9A CN112528317A (en) | 2020-11-10 | 2020-11-10 | Information processing method, device and equipment based on block chain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011244859.9A CN112528317A (en) | 2020-11-10 | 2020-11-10 | Information processing method, device and equipment based on block chain |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112528317A true CN112528317A (en) | 2021-03-19 |
Family
ID=74980097
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011244859.9A Pending CN112528317A (en) | 2020-11-10 | 2020-11-10 | Information processing method, device and equipment based on block chain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112528317A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109886683A (en) * | 2019-02-25 | 2019-06-14 | 北京神荼科技有限公司 | Monitor the method, apparatus and storage medium of block chain data |
| EP3528457A2 (en) * | 2018-02-19 | 2019-08-21 | Deutsche Telekom AG | Collaborative internet-of-things anomaly detection |
| CN110177108A (en) * | 2019-06-02 | 2019-08-27 | 四川虹微技术有限公司 | A kind of anomaly detection method, device and verifying system |
| CN110300167A (en) * | 2019-06-28 | 2019-10-01 | 京东数字科技控股有限公司 | Business information processing method, equipment and readable storage medium storing program for executing based on block chain |
| CN110533154A (en) * | 2019-08-30 | 2019-12-03 | 联想(北京)有限公司 | A kind of data processing method and electronic equipment |
| US10528890B1 (en) * | 2019-07-24 | 2020-01-07 | Kpmg Llp | Blockchain-based training data management system and method for trusted model improvements |
| CN110879827A (en) * | 2019-10-23 | 2020-03-13 | 联想(北京)有限公司 | Information processing method and device based on block chain network |
-
2020
- 2020-11-10 CN CN202011244859.9A patent/CN112528317A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3528457A2 (en) * | 2018-02-19 | 2019-08-21 | Deutsche Telekom AG | Collaborative internet-of-things anomaly detection |
| CN109886683A (en) * | 2019-02-25 | 2019-06-14 | 北京神荼科技有限公司 | Monitor the method, apparatus and storage medium of block chain data |
| CN110177108A (en) * | 2019-06-02 | 2019-08-27 | 四川虹微技术有限公司 | A kind of anomaly detection method, device and verifying system |
| CN110300167A (en) * | 2019-06-28 | 2019-10-01 | 京东数字科技控股有限公司 | Business information processing method, equipment and readable storage medium storing program for executing based on block chain |
| US10528890B1 (en) * | 2019-07-24 | 2020-01-07 | Kpmg Llp | Blockchain-based training data management system and method for trusted model improvements |
| CN110533154A (en) * | 2019-08-30 | 2019-12-03 | 联想(北京)有限公司 | A kind of data processing method and electronic equipment |
| CN110879827A (en) * | 2019-10-23 | 2020-03-13 | 联想(北京)有限公司 | Information processing method and device based on block chain network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109543075B (en) | Block generation method and system | |
| CN110110555B (en) | Voting method and device in block chain | |
| CN107220130B (en) | Method, device and system for realizing information consensus at nodes of block chain | |
| CN108711212B (en) | Voting certificate storage method, device and system | |
| Chen et al. | Robust blockchained federated learning with model validation and proof-of-stake inspired consensus | |
| WO2020258925A1 (en) | Blockchain-based service information processing method, device, and readable storage medium | |
| CN109961287A (en) | A kind of monitoring and managing method and supervisory systems of block chain | |
| CN110046896A (en) | A kind of block processes method and node and system | |
| CN110879827B (en) | Information processing method and equipment based on block chain network | |
| CN110599179B (en) | Risk control method and related equipment based on blockchain system | |
| CN109981328A (en) | A kind of fault early warning method and device | |
| CN112134762A (en) | Test method, device, terminal and storage medium for block chain network structure | |
| CN113326516A (en) | Block chain consensus method, block chain system and computer equipment | |
| CN115499379B (en) | Information interaction method, device, equipment and medium based on block chain | |
| CN114372046A (en) | Parking flow prediction model training method based on federal learning | |
| CN116233177A (en) | Block chain and dynamic reputation based Internet of vehicles data secure sharing method and system | |
| CN116192405A (en) | Electronic voting method and related device | |
| CN111865595A (en) | Block chain consensus method and device | |
| CN112418433A (en) | Model training method, device and equipment | |
| CN107705126B (en) | Transaction instruction processing method and device | |
| CN112528317A (en) | Information processing method, device and equipment based on block chain | |
| Dong et al. | Inference of online auction shills using dempster-shafer theory | |
| Panigrahi et al. | Use of dempster-shafer theory and Bayesian inferencing for fraud detection in mobile communication networks | |
| CN113596064B (en) | Analysis control method and system for security platform | |
| CN113806800B (en) | Privacy protection method and system for improving communication efficiency and reasonably distributing rewards |
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 |