CN117408332A - De-centralized AI training and transaction platform and method - Google Patents
De-centralized AI training and transaction platform and method Download PDFInfo
- Publication number
- CN117408332A CN117408332A CN202311359009.7A CN202311359009A CN117408332A CN 117408332 A CN117408332 A CN 117408332A CN 202311359009 A CN202311359009 A CN 202311359009A CN 117408332 A CN117408332 A CN 117408332A
- Authority
- CN
- China
- Prior art keywords
- model
- training
- blockchain
- federal learning
- local
- 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
- 238000012549 training Methods 0.000 title claims abstract description 139
- 238000000034 method Methods 0.000 title claims abstract description 50
- 230000002776 aggregation Effects 0.000 claims abstract description 21
- 238000004220 aggregation Methods 0.000 claims abstract description 21
- 238000012795 verification Methods 0.000 claims abstract description 10
- 230000001960 triggered effect Effects 0.000 claims abstract description 5
- 238000005516 engineering process Methods 0.000 claims description 16
- 230000007246 mechanism Effects 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 11
- 238000012360 testing method Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 238000011156 evaluation Methods 0.000 claims description 8
- 238000004422 calculation algorithm Methods 0.000 claims description 7
- 238000011161 development Methods 0.000 claims description 7
- 230000018109 developmental process Effects 0.000 claims description 7
- 238000010801 machine learning Methods 0.000 claims description 7
- 238000003058 natural language processing Methods 0.000 claims description 7
- 238000012546 transfer Methods 0.000 claims description 7
- 238000013461 design Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 230000004931 aggregating effect Effects 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 5
- 238000001303 quality assessment method Methods 0.000 claims description 5
- 238000012423 maintenance Methods 0.000 claims description 4
- 238000009826 distribution Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 10
- 230000008901 benefit Effects 0.000 description 8
- 238000007726 management method Methods 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 6
- 230000006872 improvement Effects 0.000 description 5
- 238000003745 diagnosis Methods 0.000 description 4
- 238000003860 storage Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000005284 excitation Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 238000006116 polymerization reaction Methods 0.000 description 2
- 241001122767 Theaceae Species 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
Classifications
-
- 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
- G06N3/098—Distributed learning, e.g. federated learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/04—Payment circuits
- G06Q20/06—Private payment circuits, e.g. involving electronic currency used among participants of a common payment scheme
- G06Q20/065—Private payment circuits, e.g. involving electronic currency used among participants of a common payment scheme using e-cash
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/22—Payment schemes or models
- G06Q20/223—Payment schemes or models based on the use of peer-to-peer networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/382—Payment protocols; Details thereof insuring higher security of transaction
- G06Q20/3821—Electronic credentials
- G06Q20/38215—Use of certificates or encrypted proofs of transaction rights
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q20/00—Payment architectures, schemes or protocols
- G06Q20/38—Payment protocols; Details thereof
- G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
- G06Q20/401—Transaction verification
- G06Q20/4014—Identity check for transactions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Accounting & Taxation (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- General Business, Economics & Management (AREA)
- Computer Security & Cryptography (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Technology Law (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of artificial intelligence, and discloses a decentralization AI training and transaction platform and method based on blockchain and federal learning, wherein the method comprises the following steps: the AI training layer is used for enabling the edge equipment/terminal to execute the training task of federal learning independently by utilizing local data from the blockchain side or perhaps the latest global model parameter, submitting the trained local model parameter to the blockchain layer, and enabling the training node to be added into federal learning training after identity verification by a certificate authority; the block chain layer, the federal learning training node acquires a new comprehensive global model from the block chain and starts training by using local data; after training, the federal learning training node submits the local model to the blockchain, the intelligent contract monitors the number of the local models uploaded to the blockchain in real time, and if the number reaches a first threshold value, a new round of aggregation operation is triggered; and the transaction layer is used for adding the finally generated global model into the AI model pool for any user to conduct transaction.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence, and particularly relates to a decentralization AI training and transaction platform based on blockchain and federal learning.
Background
Federal learning enables collaborative model learning without sharing raw data and is increasingly attracting attention from technology and industry that requires privacy protection. Although federal learning can solve data privacy problems between data owners, its resistance to attacks presents a significant challenge due to the frequent interactions between data and the exposure of the central server. The distributed storage in the blockchain technology is not tamper-proof, and the rights management access control technology can well solve the problems of reliable data transmission and malicious attack resistance. Therefore, blockchain techniques are widely proposed for use in federal learning to address security issues. Through investigation, there are many projects that have applied the fusion technique of blockchain and federal learning in the actual scenario. The method comprises the steps of (1) applying blockchain and federal learning to the Internet of things and an edge computing network to ensure high-efficiency and safety of the network, (2) introducing a tea variety identification platform based on blockchain and federal learning, and (3) providing a fast fashion supply chain federal sharing platform based on blockchain.
Through the above analysis, the problems and defects existing in the prior art are as follows:
1. introducing decentralised trusted edge intelligence: the artificial intelligent model aggregation process is easy to be attacked maliciously, so that the quality of the model is reduced, decentralization is completed through a block chain, malicious nodes are removed through judgment of contribution degree in the model aggregation stage, and most of edge intelligence is guaranteed to be credible.
2. Token excitation for model quality assessment: the prior art calculates the contribution degree of the participants by evaluating the precision of a local model through a test set, the evaluation mode needs to take a lot of time or calculation, and a complete contribution evaluation and incentive mechanism based on the blockchain side is needed.
3. Realize a platform based on AI ability trade: at present, the model of the federal learning side is single, the practicability is not strong, most of the existing patents and researches focus on the application in a certain aspect, and a broad federal learning model pool is not available, and the performance of various transactions is not supported. A more sophisticated transaction platform with a huge pool of AI models is needed.
Disclosure of Invention
In order to solve the problem of data privacy safety and resist malicious attacks and realize the transaction and sharing of models, the invention provides a decentralised AI training and transaction platform based on blockchain and federal learning, and the platform can provide a huge AI model pool capable of using token transactions and support federal learning training of various artificial intelligent models in different scales. The blockchain nodes and the distributed edge equipment can freely apply for model adding training after passing identity authentication, corresponding medal rewards are obtained according to the contribution degree, and the medal can be used for purchasing an AI model trained by a platform and can be sold outwards.
The invention is realized in such a way that a decentralizing AI training and transaction method based on blockchain and federal learning comprises the following steps:
1) Initializing a platform:
a. selecting an appropriate blockchain platform;
b. designing and deploying intelligent contracts to realize model training and automation of transaction process;
c. setting up a certificate authority for carrying out identity verification on the training node;
d. initializing an AI model pool and setting a transaction layer for AI model transaction;
2) Federal learning training process:
a. the training node performs identity verification through the CA and obtains a corresponding certificate;
b. the training node acquires the latest global model parameters from the blockchain side;
c. the training node executes the training task of federal learning by using the local data;
d. submitting the trained local model parameters to a blockchain layer;
3) Global model aggregation and updating:
a. in a new round of aggregation operation, all local models will be aggregated to form a new global model;
b. the new global model will be uploaded to the blockchain through the smart contracts and added to the AI model pool;
4) AI model transaction:
a. any user may select and purchase an appropriate AI model in the pool of AI models;
b. the purchased model will be labeled to indicate ownership and usage rights;
c. for the secondary use or transfer of the model, operations need to be performed through smart contracts, ensuring transfer of ownership and payment of fees;
5) Safety and performance optimization:
a. aiming at the safety requirement of the block chain platform, carrying out corresponding safety design and optimization;
b. the performance and efficiency of the platform are improved by continuously optimizing intelligent contracts and related algorithms;
6. platform maintenance and continuous iteration:
a. continuously maintaining and upgrading the platform to meet the continuously changing user requirements and technical development;
b. and according to the actual application scene and feedback, iterating and improving the platform so as to improve the user experience and performance.
Further, the polymerization operation includes:
s101, carrying out average operation on all local models of the round to obtain an average model. Respectively evaluating the deviation degree of each local model and the average model;
s102, if the deviation degree of the local model is smaller than a second threshold value, the local model is considered as a normal model, otherwise, the local model is considered as a malicious model;
s103, according to the deviation degree, assigning contribution degree of the normal model submitting node, wherein the contribution degree and the contribution degree are in inverse proportion; deducting the contribution degree of the malicious model submitted federal learning training node, and making the contribution degree and the contribution degree in direct proportion.
S104, all the normal models are aggregated, namely all the normal models are subjected to average operation, and a global model is obtained;
s105, the contribution degree of the global model and the federal learning training node of the round is uploaded into the blockchain.
Further, the degree of deviation of the local model from the average model is evaluated by:
where t represents the number of rounds, k represents the number of local models,representing the similarity of the two models of the t-th round, < >>Weights representing the t-th round average model, +.>Weights representing the kth local model of the t-th round, +.>Representing Min Shi distance of the t-th local model from the t-th average model.
If at least 50% of the nodes in the intra-chip blockchain network are honest, the second threshold value is s_mid+Aσ; where s_mid is the median of the deviation set, σ is the standard deviation of the deviation set distribution, and a is an adjustable constant coefficient.
When the accumulated contribution of a node is below a third threshold, the smart contract exercises access control to disqualify the node for participation in all system transactions.
Further, a token incentive mechanism for model quality assessment, comprising:
the federal learning can evaluate the contribution degree of the participated training party according to individual methods, leave-one-out methods, xia Puli values, minimum kernels and other methods; the blockchain evaluates according to the contribution degree of the completion of the federal learning and rewards and punishs tokens according to the size of the contribution degree.
Further, the operation flow of the decentralizing AI training and transaction platform based on blockchain and federal learning comprises the following steps:
s201, edge equipment in a distributed AI training layer performs trusted federation learning based on block chains;
s202, the blockchain carries out contribution evaluation and medal issuing according to an incentive mechanism in an intelligent contract according to a model submitted by edge equipment participating in federal learning tasks, and a user participating in federal learning training can obtain corresponding medals as incentives;
s203, the blockchain node and the edge equipment are added with training after identity authentication is carried out on a certificate authority;
s204, the global model on the blockchain forms a huge shared AI model pool, wherein the AI model pool comprises AI models such as intelligent security, intelligent power grid, medical image processing, natural language processing and the like, and a purchaser can use tokens to conduct transactions of the AI models. Tokens may be used to purchase platform-trained AI models, or may be sold externally.
Another object of the present invention is to provide a decentralised AI training and transaction method based on blockchain and federal learning, comprising:
s1, a task publisher publishes training tasks, wherein the training tasks comprise: intelligent security, smart grid, medical image processing, natural language processing, and the like, and provides a test set and an untrained machine learning model as an initial global model.
S2, the nodes in the block chain network download an untrained machine learning model, and initialize parameters of the model for training.
S3, the federal learning training node trains the model by utilizing a local user data set on the basis of the latest comprehensive global model (or initial model). And obtaining a local model after training is completed. The participants initiate a proposal request (i.e., transaction request) for the uplink and upload trained local model parameters to the blockchain.
S4, when the number of the local models in the blockchain accords with the aggregation condition, the intelligent contract submits a request for aggregating the model parameters, generates a new global model and tests the precision of the global model by using a test set.
S5, the finally generated global model is added into an AI model pool to be traded by any user.
It is another object of the present invention to provide a computer device including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the blockchain and federal learning based decentralizing AI training and transaction method.
It is another object of the present invention to provide a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the blockchain and federal learning based de-centralized AI training and transaction method.
The invention further aims to provide an information data processing terminal which is used for realizing the decentralization AI training and transaction platform based on block chain and federation learning.
Another object of the present invention is to provide a decentralised AI training and trading platform based on blockchain and federal learning, comprising:
the AI training layer is used for enabling the edge equipment/terminal to execute the training task of federal learning independently by utilizing local data from the blockchain side or perhaps the latest global model parameter, submitting the trained local model parameter to the blockchain layer, and enabling the training node to be added into federal learning training after identity verification by a certificate authority;
the block chain layer, the federal learning training node acquires a new comprehensive global model from the block chain and starts training by using local data; after training, the federal learning training node submits the local model to the blockchain, the intelligent contract monitors the number of the local models uploaded to the blockchain in real time, and if the number reaches a first threshold value, a new round of aggregation operation is triggered;
and the transaction layer is used for adding the finally generated global model into the AI model pool for any user to conduct transaction.
In combination with the technical scheme and the technical problems to be solved, the technical scheme to be protected has the following advantages and positive effects:
first, traditional machine learning methods require centralized data training, which may involve sending sensitive user data to a central server, with potential privacy risks. Federal learning proposes the concept of decentralized training, but has the difficulty of protecting the privacy of user data while effectively coordinating the training of global models. The present invention creates a global model by moving model training to the users' locality, each user training the local model while maintaining the data locality, and then synthesizing these local models using smart contracts and parameter aggregation techniques. The contribution degree of the user node is determined by measuring the deviation degree between the local model and the global model, so that the method not only solves the problem of data privacy, but also effectively coordinates the training of the global model, and provides creative privacy security effects.
Blockchain technology typically involves consensus algorithms and smart contracts, but when combined with AI training, it is necessary to address the issues of parameter aggregation and node security. Parameter aggregation needs to ensure that data does not leak or be tampered with when the blockchain layer merges model parameters. In addition, the problem of malicious node attack needs to be solved to protect the security of the whole system. According to the invention, through designing the intelligent contract and the consensus algorithm, the safe aggregation of parameters is realized, and malicious nodes can be removed. In addition, nodes are encouraged to engage in good behavior by evaluating contributions and issuing token incentives based on the contribution of the user nodes. The combination technology not only ensures the safety of the block chain layer, but also protects the overall safety of the distributed AI training, and has creative node safety effect.
The present invention provides a novel AI model trading platform in which a user can trade a trained AI model, but this involves how to ensure the legitimacy of the trade and the privacy of the model. The smart contracts must approve the transaction and ensure confidentiality of the model. The invention provides a safe and traceable model transaction mode by examining and approving model transaction through intelligent contracts in a blockchain and taking the medal as the cost of the transaction. The technology not only supports model sharing, but also protects the privacy of the model, and brings creative technical effects to the field of AI model transaction.
The invention successfully solves a plurality of technical problems related to federal learning, blockchain, data privacy, AI model transaction and the like, and brings a series of creative technical effects for the fields, including data privacy protection, node security and model transaction innovation. These effects promote the development of distributed AI while protecting users and their data, with significant technical and commercial potential.
Secondly, the invention successfully solves the problem of single-point fault existing in the traditional edge intelligent computing by introducing a decentralization architecture. The architecture not only improves the robustness and reliability of the system, but also protects the user data and privacy. The user data is processed and trained locally and does not leave the local, so that potential privacy risks are reduced, and higher security and reliability are provided for intelligent edge calculation.
The invention introduces a model quality assessment and incentive mechanism, and encourages edge computing devices to provide high quality models through intelligent contract incentive algorithms. This mechanism ensures the accuracy and reliability of the edge smart computation model, while inspiring the enthusiasm of users, making them more willing to participate in the contribution of the model. This helps to improve the performance and efficiency of the overall system.
The present invention designs a transaction layer and a huge pool of AI models, allowing any user to conduct AI model transactions using virtual tokens. The design not only improves the sharing and usability of the AI model, but also ensures the safety and reliability of the transaction through the blockchain authorization. Users can exchange models with confidence without worrying about data leakage or model integrity issues, thereby promoting the development of AI model transactions.
The invention creates a transaction platform based on AI capability, and applies a blockchain and federal learning fusion technology to an actual scene. The existence of the platform enables more users to participate in training and trading of the artificial intelligence model, and promotes the wide spread and utilization of AI technology. The user can select and exchange the model according to own requirements, so that personalized AI application is realized, and innovation and popularization of AI technology are promoted.
In summary, the technical scheme of the invention not only solves the key problems in the existing edge intelligent computing, but also brings numerous innovative technical effects and advantages, including data privacy protection, model quality improvement, transaction safety and AI technology popularization. These effects and advantages make the present invention of significant commercial and social significance in the field of edge intelligent computing and AI model trading.
Thirdly, as inventive supplementary evidence of the claims of the present invention, the following important aspects are also presented:
(1) The expected benefits and commercial values after the technical scheme of the invention is converted are as follows:
the technical scheme of the invention has great commercial potential and expected benefits. Firstly, by providing a safe edge intelligent computing and AI model transaction platform, the invention provides an efficient way for enterprises to utilize the distributed AI resources, reduces the training and reasoning cost of the AI model, and improves the intelligent level of decision making and application. The method brings great benefit in the field of enterprises, and improves production efficiency and competitiveness. And secondly, the technology of the invention can be also used in the fields of medical diagnosis, intelligent cities, automatic driving and the like, has wide application prospect and is expected to promote the development of AI technology in the global scope. Thus, the commercial value and expected return of the present invention is considerable.
(2) The technical scheme of the invention fills the technical blank in the domestic and foreign industries:
in the fields of edge intelligent computing and AI model transaction, a plurality of technical problems of data privacy protection, node security, model sharing and the like exist. The invention provides a comprehensive solution by fusing federal learning and blockchain technologies, and fills the technical blank in the fields. The technical level in the field is relatively lagging in the domestic and foreign industries, and the innovative technical scheme of the invention leads the field to be in the leading position.
(3) Whether the technical scheme of the invention solves the technical problems that people want to solve all the time but fail to obtain success all the time is solved:
the technical scheme of the invention solves the important problems which have long plagued the AI field, namely how to protect the privacy of user data in a distributed environment, how to ensure the quality and safety of a model and how to establish a trusted model transaction platform. These problems have been a hotspot concern in academia and industry, but have long failed to be a successful solution. The invention successfully solves the technical problems by innovative thinking and comprehensive technical scheme.
(4) The technical scheme of the invention overcomes the technical bias:
the invention overcomes the technical bias of the traditional centralized AI model training and transaction method. The traditional method relies on centralized data storage and processing, and has the concerns of data privacy and safety, and the invention breaks through the prejudice through the technical means of decentralization, blockchain, federal learning and the like, realizes high-efficiency distributed AI model training and transaction, and provides safer and more credible AI services for users. The technical solution of the present invention is therefore an important challenge and improvement of the conventional technical prejudice.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a decentralised AI training and trading platform based on blockchain and federal learning in accordance with an embodiment of the present invention;
FIG. 2 is a flow chart of the operation of the distributed AI training layer provided by an embodiment of the invention;
FIG. 3 is a flow chart of the operation of the blockchain layer provided by an embodiment of the present invention;
FIG. 4 is a flow chart of the operation of the transaction layer provided by an embodiment of the present invention;
FIG. 5 is a schematic view of a front end page on a vehicle according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a hardware module and a specific implementation scenario provided in an embodiment of the present invention;
FIG. 7 is an accuracy evaluation chart of eye and mouth models in fatigue driving detection provided by an embodiment of the invention; wherein, (a) eyes, (b) mouth.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Aiming at the problems existing in the prior art, the invention provides a decentralised AI training and transaction platform based on blockchain and federal learning, and the invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the de-centralized AI training and transaction platform based on blockchain and federal learning provided by the embodiment of the present invention includes:
as shown in fig. 2, the AI training layer is configured to enable the edge device/terminal to perform a training task of federal learning independently by using local data, submit the trained local model parameters to the blockchain layer, and enable the training node to join federal learning training after identity verification by a certificate authority;
the block chain layer, the federal learning training node acquires a new comprehensive global model from the block chain and starts training by using local data; after training, the federal learning training node submits the local model to the blockchain, the intelligent contract monitors the number of the local models uploaded to the blockchain in real time, and if the number reaches a first threshold value, a new round of aggregation operation is triggered;
and the transaction layer is used for adding the finally generated global model into the AI model pool for any user to conduct transaction.
The polymerization operation includes:
s101, carrying out average operation on all local models of the round to obtain an average model. Respectively evaluating the deviation degree of each local model and the average model;
s102, if the deviation degree of the local model is smaller than a second threshold value, the local model is considered as a normal model, otherwise, the local model is considered as a malicious model;
s103, according to the deviation degree, assigning contribution degree of the normal model submitting node, wherein the contribution degree and the contribution degree are in inverse proportion; deducting the contribution degree of the malicious model submitted federal learning training node, and making the contribution degree and the contribution degree in direct proportion.
S104, all the normal models are aggregated, namely all the normal models are subjected to average operation, and a global model is obtained;
s105, the contribution degree of the global model and the federal learning training node of the round is uploaded into the blockchain.
The degree of deviation of the local model from the average model is evaluated by:
representing the similarity of the two models of the t-th round, < >>Weights representing the t-th round average model, +.>Weights representing the kth local model of the t-th round, +.>Min Shi distance representing the t-th local model and the t-th average model; if at least 50% of the nodes in the intra-chip blockchain network are honest, the second threshold value is s_mid+Aσ; where s_mid is the median of the deviation set, σ is the standard deviation of the deviation set distribution, and A is an adjustable constant coefficient; when the accumulated contribution of a node is below a third threshold, the smart contract exercises access control to disqualify the node for participation in all system transactions.
A token incentive mechanism for model quality assessment, comprising:
the federal learning can evaluate the contribution degree of the participated training party according to individual methods, leave-one-out methods, xia Puli values, minimum kernels and other methods; the blockchain evaluates according to the contribution degree of the completion of the federal learning and rewards and punishs tokens according to the size of the contribution degree.
The operation flow of the decentralization AI training and transaction platform based on blockchain and federal learning comprises the following steps:
s201, edge equipment in a distributed AI training layer performs trusted federation learning based on block chains;
s202, the blockchain carries out contribution evaluation and medal issuing according to an incentive mechanism in an intelligent contract according to a model submitted by edge equipment participating in federal learning tasks, and a user participating in federal learning training can obtain corresponding medals as incentives;
s203, the blockchain node and the edge equipment are added with training after identity authentication is carried out on a certificate authority;
s204, the global model on the blockchain forms a huge shared AI model pool, wherein the AI model pool comprises AI models such as intelligent security, intelligent power grid, medical image processing, natural language processing and the like, and a purchaser can use tokens to conduct transactions of the AI models. Tokens may be used to purchase platform-trained AI models, or may be sold externally.
The embodiment of the invention provides a decentralizing AI training and trading method based on blockchain and federal learning, which comprises the following steps:
s1, a task publisher publishes training tasks, wherein the training tasks comprise: intelligent security, smart grid, medical image processing, natural language processing, and the like, and provides a test set and an untrained machine learning model as an initial global model.
S2, the nodes in the block chain network download an untrained machine learning model, and initialize parameters of the model for training.
S3, the federal learning training node trains the model by utilizing a local user data set on the basis of the latest comprehensive global model (or initial model). And obtaining a local model after training is completed. The participants initiate a proposal request (i.e., transaction request) for the uplink and upload trained local model parameters to the blockchain.
S4, when the number of the local models in the blockchain accords with the aggregation condition, the intelligent contract submits a request for aggregating the model parameters, generates a new global model and tests the precision of the global model by using a test set.
S5, the finally generated global model is added into an AI model pool to be traded by any user.
The blockchain judges malicious models by evaluating the deviation degree of each local model from the average model and assigns the contribution degree of each edge device training node according to an excitation mechanism. And removing the malicious models, aggregating all normal models by the nodes to obtain a global model of the round, and submitting the global model to a chain. After the uplink is submitted, the training node is stimulated to issue tokens according to a stimulation mechanism in the intelligent contract. The blockchain node also needs to be authenticated by a certificate authority to join the blockchain network. In addition, the blockchain has the characteristic of evidence storage and tracing, namely, records added into the blockchain are permanently stored, trader information is bound in each transaction record in the blockchain, and the records are completely recorded and traced and cannot be destroyed or tampered. The operational flow of the blockchain layer is shown in fig. 3.
At the transaction level, any user can purchase platform-trained AI models with tokens, which are located on the blockchain, or sell AI models on his own hand to obtain tokens. The AI model pool contains all trained global models on the blockchain chain, such as artificial intelligent models of intelligent security, intelligent power grid, medical image processing, natural language processing and the like, so that the transaction user can conveniently purchase and sell the models. When a user needs to conduct a transaction, firstly, an application is submitted to the blockchain, and after the blockchain approval passes, the transaction can be conducted. The operational flow of the transaction layer is shown in fig. 4.
An application embodiment of the present invention provides a computer device including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of a decentralised AI training and transaction method based on blockchain and federal learning.
An application embodiment of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of a decentralised AI training and transaction method based on blockchain and federal learning.
The embodiment of the application of the invention provides an information data processing terminal which is used for realizing a decentralization AI training and transaction platform based on block chain and federal learning.
The following is a method for constructing a decentralised AI training and transaction platform based on blockchain and federal learning:
1. platform initialization:
a. select the appropriate blockchain platform (e.g., ethernet, corda, etc.) and set the corresponding node.
b. Intelligent contracts are designed and deployed for automating model training and transaction processes.
c. A Certificate Authority (CA) is set up for authentication of the training node.
d. An AI model pool is initialized and a transaction layer is set for AI model transactions.
2. Federal learning training process:
a. the training node performs identity verification through the CA and obtains a corresponding certificate.
b. The training node obtains the latest global model parameters from the blockchain side.
c. The training node executes the training task of federal learning by using the local data and submits the trained local model parameters to the blockchain layer.
d. The intelligent contract monitors the number of local models uploaded into the blockchain in real time, and if the number reaches a first threshold value, a new round of aggregation operation is triggered.
3. Global model aggregation and updating:
a. in a new round of aggregation operation, all local models will be aggregated to form a new global model.
b. The new global model will be uploaded to the blockchain through the smart contracts and added to the AI model pool.
Ai model transaction:
a. any user may select and purchase an appropriate AI model in the pool of AI models.
b. The purchased model will be labeled to indicate ownership and usage rights.
c. For secondary use or transfer of the model, operations need to be performed through smart contracts to ensure transfer of ownership and payment of fees.
5. Safety and performance optimization:
a. and aiming at the safety requirement of the block chain platform, carrying out corresponding safety design and optimization.
b. By continuously optimizing the intelligent contracts and related algorithms, the performance and efficiency of the platform are improved.
6. Platform maintenance and continuous iteration:
a. the platform is continuously maintained and upgraded to meet changing user requirements and technological developments.
b. And according to the actual application scene and feedback, iterating and improving the platform so as to improve the user experience and performance.
Example 1: intelligent medical diagnostic application
The technical scheme of the invention can be applied to the field of intelligent medical diagnosis. Hospitals and medical facilities can use the decentralized edge intelligence architecture provided by the present invention to build a secure medical diagnostic system. The patient's medical data can be kept local and not leave the hospital network, thereby protecting the patient's privacy. Each hospital can independently train local medical models and then aggregate the models using blockchain techniques to form a global medical diagnostic model.
The system may also use model quality based assessment and incentive mechanisms to encourage hospitals to provide high quality medical models. When the medical model is excellent in diagnosing patients, hospitals can obtain token incentives, thereby improving the accuracy and reliability of medical diagnosis.
Example 2: intelligent urban traffic management
In the field of intelligent cities, the technical scheme of the invention can be used for traffic management. The traffic monitoring cameras, sensors and other devices can collect and process data locally without transmitting images and sensor data to a central server. Each traffic monitoring point may train a local model to identify traffic violations, congestion conditions, and the like.
By using the blockchain technology, the models of all traffic monitoring points can be safely aggregated to form a traffic management system in the whole city. Model quality based assessment and incentive mechanisms may encourage individual monitoring points to provide accurate traffic data and models to obtain token incentives.
Effect 1: data privacy protection
In the edge intelligent computing system implementing the invention, the data privacy is fully protected. In contrast to traditional centralising methods, the user's data does not need to be transmitted to a central server, but rather is handled and trained locally. This advantage is particularly evident in medical diagnostic applications. In a medical diagnosis system, the technical scheme of the invention is used for greatly reducing the number of data transmission and remarkably reducing the risk of privacy disclosure. In addition, through data verification and authority management of the blockchain, only authorized users can access the data, and privacy protection effect is further enhanced.
Effect 2: model quality improvement
The evaluation and excitation mechanism based on the model quality is verified in the embodiment, and the improvement of the model quality is significantly influenced. In experiments in traffic management systems, the model quality of individual traffic monitoring points was improved over a period of time. This is because the monitoring points are actively engaged in the training and contribution of the model to acquire the token rewards, thereby improving the accuracy of the model.
It should be noted that the embodiments of the present invention can be realized in hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those of ordinary skill in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The device of the present invention and its modules may be implemented by hardware circuitry, such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., as well as software executed by various types of processors, or by a combination of the above hardware circuitry and software, such as firmware. By applying the technology in the scene of fatigue driving detection of the Internet of vehicles, the front end page of the vehicle is shown in fig. 5, the hardware module and the specific implementation scene are shown in fig. 6, and in the fatigue driving detection, the accuracy evaluation diagrams of the eye and mouth models are shown in fig. 7 (a) and (b), so that the accuracy of the models can still be kept at a higher level under the condition of malicious attack.
The foregoing is merely illustrative of specific embodiments of the present invention, and the scope of the invention is not limited thereto, but any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention will be apparent to those skilled in the art within the scope of the present invention.
Claims (10)
1. A decentralizing AI training and trading method based on blockchain and federal learning is characterized in that a decentralizing AI training and trading platform is constructed by using a blockchain technology and a federal learning algorithm; the platform realizes the functions of identity verification of training nodes, global aggregation and updating of the AI model, purchase and ownership transfer of the AI model and the like through the intelligent contract automated training and transaction process of the AI model. At the same time, the platform performs continuous maintenance and iteration to accommodate changing user needs and technological developments.
2. The blockchain and federal learning-based de-centralized AI training and transaction method of claim 1, comprising:
1) Initializing a platform:
a. selecting an appropriate blockchain platform;
b. designing and deploying intelligent contracts to realize model training and automation of transaction process;
c. setting up a certificate authority for carrying out identity verification on the training node;
d. initializing an AI model pool and setting a transaction layer for AI model transaction;
2) Federal learning training process:
a. the training node performs identity verification through the CA and obtains a corresponding certificate;
b. the training node acquires the latest global model parameters from the blockchain side;
c. the training node executes the training task of federal learning by using the local data;
d. submitting the trained local model parameters to a blockchain layer;
3) Global model aggregation and updating:
a. in a new round of aggregation operation, all local models will be aggregated to form a new global model;
b. the new global model will be uploaded to the blockchain through the smart contracts and added to the AI model pool;
4) AI model transaction:
a. any user may select and purchase an appropriate AI model in the pool of AI models;
b. the purchased model will be labeled to indicate ownership and usage rights;
c. for the secondary use or transfer of the model, operations need to be performed through smart contracts, ensuring transfer of ownership and payment of fees;
5) Safety and performance optimization:
a. aiming at the safety requirement of the block chain platform, carrying out corresponding safety design and optimization;
b. the performance and efficiency of the platform are improved by continuously optimizing intelligent contracts and related algorithms;
6) Platform maintenance and continuous iteration:
a. continuously maintaining and upgrading the platform to meet the continuously changing user requirements and technical development;
b. and according to the actual application scene and feedback, iterating and improving the platform so as to improve the user experience and performance.
3. The blockchain and federal learning-based de-centralized AI training and transaction method of claim 2, wherein the aggregating operation comprises:
s101, carrying out average operation on all local models of the round to obtain an average model; respectively evaluating the deviation degree of each local model and the average model;
s102, if the deviation degree of the local model is smaller than a second threshold value, the local model is considered as a normal model, otherwise, the local model is considered as a malicious model;
s103, according to the deviation degree, assigning contribution degree of the normal model submitting node, wherein the contribution degree and the contribution degree are in inverse proportion; deducting contribution degree of the malicious model submitted federal learning training node, wherein the contribution degree and the contribution degree are in direct proportion;
s104, all the normal models are aggregated, namely all the normal models are subjected to average operation, and a global model is obtained;
s105, the contribution degree of the global model and the federal learning training node of the round is uploaded into the blockchain.
4. The blockchain and federal learning-based de-centralized AI training and transaction method of claim 3, wherein the degree of deviation of the local model from the average model is assessed by:
representing the similarity of the two models of the t-th round, < >>Weights representing the t-th round average model, +.>Weights representing the kth local model of the t-th round, +.>Min Shi distance representing the t-th local model and the t-th average model; if at least 50% of the nodes in the intra-chip blockchain network are honest, the second threshold value is s_mid+Aσ; where s_mid is the median of the deviation set, σ is the standard deviation of the deviation set distribution, and A is an adjustable constant coefficient; when the accumulated contribution of a node is below a third threshold, the smart contract exercises access control to disqualify the node for participation in all system transactions.
5. The blockchain and federal learning-based de-centralized AI training and transaction method of claim 3, wherein the token incentive mechanism for model quality assessment comprises:
the federal learning can evaluate the contribution degree of the participated training party according to individual methods, leave-one-out methods, xia Puli values, minimum kernels and other methods; the blockchain evaluates according to the contribution degree of the completion of the federal learning and rewards and punishs tokens according to the size of the contribution degree.
6. The blockchain and federal learning-based decentralizing AI training and transaction method of claim 3, wherein the blockchain and federal learning-based decentralizing AI training and transaction method operational flow includes:
s201, edge equipment in a distributed AI training layer performs trusted federation learning based on block chains;
s202, the blockchain carries out contribution evaluation and medal issuing according to an incentive mechanism in an intelligent contract according to a model submitted by edge equipment participating in federal learning tasks, and a user participating in federal learning training can obtain corresponding medals as incentives;
s203, the blockchain node and the edge equipment are added with training after identity authentication is carried out on a certificate authority;
s204, the global model on the blockchain forms a huge shared AI model pool, wherein the AI model pool comprises AI models such as intelligent security, intelligent power grid, medical image processing, natural language processing and the like, and a purchaser can use medals to conduct transactions of the AI models; tokens may be used to purchase platform-trained AI models, or may be sold externally.
7. A method for decentralizing AI training and trading based on blockchain and federal learning, comprising:
s1, a task publisher publishes training tasks, wherein the training tasks comprise: intelligent security, smart grid, medical image processing, natural language processing, and providing a test set and an untrained machine learning model as an initial global model;
s2, downloading an untrained machine learning model by a node in the blockchain network, and initializing parameters of the model for training;
s3, the federal learning training node trains a model by utilizing a local user data set on the basis of the latest comprehensive global model, and a local model is obtained after training is completed; the participants initiate a proposal request of the uplink, and the trained local model parameters are uploaded to the blockchain;
s4, when the number of the local models in the blockchain accords with the aggregation condition, the intelligent contract submits a request for aggregating the model parameters, generates a new global model and tests the precision by using a test set;
s5, the finally generated global model is added into an AI model pool to be traded by any user.
8. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the blockchain and federal learning based de-centralized AI training and transaction method of claim 7.
9. An information data processing terminal for implementing the decentralised AI training and transaction platform based on blockchain and federal learning as claimed in any one of claims 1 to 6.
10. A decentralised AI training and trading platform based on blockchain and federal learning, comprising:
the AI training layer is used for enabling the edge equipment/terminal to execute the training task of federal learning independently by utilizing local data from the blockchain side or perhaps the latest global model parameter, submitting the trained local model parameter to the blockchain layer, and enabling the training node to be added into federal learning training after identity verification by a certificate authority;
the block chain layer, the federal learning training node acquires a new comprehensive global model from the block chain and starts training by using local data; after training, the federal learning training node submits the local model to the blockchain, the intelligent contract monitors the number of the local models uploaded to the blockchain in real time, and if the number reaches a first threshold value, a new round of aggregation operation is triggered;
and the transaction layer is used for adding the finally generated global model into the AI model pool for any user to conduct transaction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311359009.7A CN117408332A (en) | 2023-10-19 | 2023-10-19 | De-centralized AI training and transaction platform and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311359009.7A CN117408332A (en) | 2023-10-19 | 2023-10-19 | De-centralized AI training and transaction platform and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117408332A true CN117408332A (en) | 2024-01-16 |
Family
ID=89488506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311359009.7A Pending CN117408332A (en) | 2023-10-19 | 2023-10-19 | De-centralized AI training and transaction platform and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117408332A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112329940A (en) * | 2020-11-02 | 2021-02-05 | 北京邮电大学 | Personalized model training method and system combining federal learning and user portrait |
| CN113420335A (en) * | 2021-08-24 | 2021-09-21 | 浙江数秦科技有限公司 | Block chain-based federal learning system |
| CN114186237A (en) * | 2021-10-26 | 2022-03-15 | 北京理工大学 | Truth-value discovery-based robust federated learning model aggregation method |
| CN114448601A (en) * | 2022-03-01 | 2022-05-06 | 南京航空航天大学 | Distributed federal learning security defense method based on end edge cloud architecture and application |
| CN115081002A (en) * | 2022-06-28 | 2022-09-20 | 西安电子科技大学 | Aggregation server selection method for decentralized federal learning |
| CN115271089A (en) * | 2022-06-15 | 2022-11-01 | 京信数据科技有限公司 | Block chain-based federal learning credible training method and device |
| CN115834409A (en) * | 2022-11-22 | 2023-03-21 | 上海交通大学 | Federal learning-oriented safety aggregation method and system |
| CN115907029A (en) * | 2022-11-08 | 2023-04-04 | 北京交通大学 | Defense method and system for federal learning virus attack |
| CN116627970A (en) * | 2023-05-18 | 2023-08-22 | 浙江大学 | Data sharing method and device based on blockchain and federal learning |
-
2023
- 2023-10-19 CN CN202311359009.7A patent/CN117408332A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112329940A (en) * | 2020-11-02 | 2021-02-05 | 北京邮电大学 | Personalized model training method and system combining federal learning and user portrait |
| CN113420335A (en) * | 2021-08-24 | 2021-09-21 | 浙江数秦科技有限公司 | Block chain-based federal learning system |
| CN114186237A (en) * | 2021-10-26 | 2022-03-15 | 北京理工大学 | Truth-value discovery-based robust federated learning model aggregation method |
| CN114448601A (en) * | 2022-03-01 | 2022-05-06 | 南京航空航天大学 | Distributed federal learning security defense method based on end edge cloud architecture and application |
| CN115271089A (en) * | 2022-06-15 | 2022-11-01 | 京信数据科技有限公司 | Block chain-based federal learning credible training method and device |
| CN115081002A (en) * | 2022-06-28 | 2022-09-20 | 西安电子科技大学 | Aggregation server selection method for decentralized federal learning |
| CN115907029A (en) * | 2022-11-08 | 2023-04-04 | 北京交通大学 | Defense method and system for federal learning virus attack |
| CN115834409A (en) * | 2022-11-22 | 2023-03-21 | 上海交通大学 | Federal learning-oriented safety aggregation method and system |
| CN116627970A (en) * | 2023-05-18 | 2023-08-22 | 浙江大学 | Data sharing method and device based on blockchain and federal learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liao et al. | Digital twin consensus for blockchain-enabled intelligent transportation systems in smart cities | |
| CN109472596B (en) | Alliance's chain common recognition method and system based on transaction assessment | |
| CN110766406B (en) | Resource transfer method, resource transfer device, storage medium and electronic equipment | |
| JP2024500137A (en) | Marketplace orchestration system to facilitate electronic market transactions | |
| CN112434280A (en) | Block chain-based federal learning defense method | |
| Hu et al. | A blockchain-based secure transaction model for distributed energy in Industrial Internet of Things | |
| JP6721903B2 (en) | Data transmitting/receiving method, data transmitting/receiving system, processing device, and computer program | |
| CN112685766A (en) | Enterprise credit investigation management method and device based on block chain, computer equipment and storage medium | |
| CN112465627A (en) | Financial loan auditing method and system based on block chain and machine learning | |
| CN116385164B (en) | Block chain-based carbon asset transaction system and method | |
| CN114048515B (en) | Medical big data sharing method based on federal learning and block chain | |
| US20230139137A1 (en) | Tokenized carbon credit trading platform | |
| CN111368262A (en) | Artificial intelligence model protection and loose coupling distributed training method based on block chain | |
| CN112184428A (en) | Processing method and system for operation and flow of financing requester | |
| CN117235692B (en) | Web 3-based software supply chain risk management system, method and platform | |
| CN114066584A (en) | Method and device for risk prevention and control of block chain | |
| CN110599370A (en) | Property management method and device based on block chain and storage medium | |
| Asamoah et al. | A blockchain-based crowdsourcing loan platform for funding higher education in developing countries | |
| CN109064329A (en) | A kind of calculation power method of commerce and calculate channel | |
| CN112200637A (en) | Financing lease transaction processing method and system based on block chain | |
| Bhardwaj et al. | Applications of blockchain in various domains | |
| CN117408332A (en) | De-centralized AI training and transaction platform and method | |
| Ali et al. | A systematic review of federated learning incentive mechanisms and associated security challenges | |
| Guo et al. | Reputation-based blockchain for spatial crowdsourcing in vehicular networks | |
| CN112702410B (en) | Evaluation system, method and related equipment based on blockchain network |
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 |