CN111858756A - Processing method, node and medium for AI training task based on block chain - Google Patents
Processing method, node and medium for AI training task based on block chain Download PDFInfo
- Publication number
- CN111858756A CN111858756A CN202010624145.4A CN202010624145A CN111858756A CN 111858756 A CN111858756 A CN 111858756A CN 202010624145 A CN202010624145 A CN 202010624145A CN 111858756 A CN111858756 A CN 111858756A
- Authority
- CN
- China
- Prior art keywords
- training
- result
- node
- model data
- detection sample
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the invention provides a processing method, a node and a medium for an AI training task based on a block chain, which are characterized in that a first training result detection sample and a training result reward scheme are issued to a block chain network, model data issued by an AI training node are received, and a second calculation result corresponding to the first training result detection sample is issued in the block chain network after the model data is used for the first calculation result of the first training result detection sample, so that a billing node in the block chain network can obtain the accuracy of the model data by comparing the first calculation result with the second calculation result, and the reward amount to the AI training node is recorded according to the accuracy and the training result reward scheme, thereby not only realizing the AI training task processing based on the block chain network, but also reducing the calculation amount of the AI training node in the model accuracy verification process, the computational pressure of the accounting node is reduced.
Description
Technical Field
The embodiment of the invention relates to the technical field of block chains, in particular to a processing method, a node and a medium for an AI training task based on a block chain.
Background
With the continuous development of Artificial Intelligence (AI) technology, AI technology can be applied in many fields, for example, in model training such as speech recognition and machine translation.
Typically, one AI training requires a large amount of computation. For example, one model training for speech recognition may involve floating point calculations for 20E Flops, and one model training for machine translation may involve floating point calculations for 10E Flops or even hundreds of EFlops. Often a very large computer cluster needs to be built in order to complete an AI training. The cost of building large-scale computer clusters is high.
Disclosure of Invention
The embodiment of the invention provides a processing method, a node and a medium for an AI training task based on a block chain, which are used for processing the AI training task through the block chain and reducing the pressure of an accounting node.
A first aspect of an embodiment of the present invention provides a method for processing an AI training task based on a block chain, including:
an AI training task initiating node issues an AI training task in a blockchain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample; the AI training task initiating node receives model data issued by the AI training node based on the AI training task and a first calculation result of the model data aiming at the first training result detection sample; and the AI training task initiating node issues a second calculation result corresponding to the first training result detection sample in the blockchain network so that the accounting node in the blockchain network obtains the accuracy of the model data by comparing the first calculation result with the second calculation result, and records the reward amount of the AI training node according to the training result reward scheme and the accuracy.
Optionally, after the AI training task initiating node receives model data issued by the AI training node based on the AI training task and a first calculation result of the model data for the first training result detection sample, the method further includes:
and the AI training task initiating node issues a second training result detection sample in the blockchain network so that the accounting node verifies the accuracy of the model data based on the second training result detection sample, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
A second aspect of the embodiments of the present invention provides a method for processing an AI training task based on a block chain, including:
the method comprises the steps that an accounting node receives an AI training task issued by an AI training task initiating node in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to a sample; the accounting node receives model data issued by an AI training node according to the AI training task and a first calculation result of the model data aiming at the first training result detection sample; the accounting node receives a second calculation result of the first training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result; and the accounting node compares the first calculation result with the second calculation result to obtain the accuracy of the model data, and records the reward amount of the AI training node according to the training result reward scheme and the accuracy.
Optionally, the method further includes:
and the accounting node receives a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
Optionally, after the accounting node receives a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, the method further includes:
and the accounting node verifies the accuracy of the model data based on the second training result detection sample.
A third aspect of the embodiments of the present invention provides an AI training task initiating node, including:
a processor and a memory, the memory having instructions stored therein that when executed by the processor perform the following:
issuing an AI training task in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result reward scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample; receiving model data issued by an AI training node based on the AI training task and a first calculation result of the model data aiming at the first training result detection sample; and issuing a second calculation result corresponding to the first training result detection sample in the blockchain network so that the accounting node in the blockchain network obtains the accuracy of the model data by comparing the first calculation result with the second calculation result, and recording the reward amount of the AI training node according to the training result reward scheme and the accuracy.
Optionally, the processor is further configured to:
and issuing a second training result detection sample in the blockchain network so that the accounting node verifies the accuracy of the model data based on the second training result detection sample, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
A fourth aspect of the present invention provides an accounting node, including: a processor and a memory, the memory having instructions stored therein that when executed by the processor perform the following:
receiving an AI training task issued by an AI training task initiating node in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample; receiving model data issued by an AI training node according to the AI training task and a first calculation result of the model data aiming at the first training result detection sample; receiving a second calculation result of the first training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result; and comparing the first calculation result with the second calculation result to obtain the accuracy of the model data, and recording the reward amount of the AI training node according to the training result reward scheme and the accuracy.
Optionally, the processor is further configured to:
and receiving a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
A fifth aspect of embodiments of the present invention provides a computer-readable storage medium, comprising instructions that, when executed on a computer, cause the computer to perform the method of the first or second aspect.
In the embodiment of the invention, the AI training task initiating node issues an AI training task carrying a first training result detection sample and a training result reward scheme into the blockchain network, and issues a second calculation result corresponding to the first training result detection sample in the blockchain network after receiving model data issued by the AI training node based on the AI training task and a first calculation result of the model data aiming at the first training result detection sample, so that the accounting node in the blockchain network can obtain the accuracy of the model data by comparing the first calculation result with the second calculation result, and record the reward amount to the AI training node according to the accuracy and the training result reward scheme, thereby not only realizing the AI training task processing based on the blockchain network, but also reducing the calculation amount of the accounting node in the model accuracy verification process, the computational pressure of the accounting node is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 is a schematic structural diagram of a communication system according to an embodiment of the present invention;
FIG. 2 is a flowchart of a block chain-based artificial intelligence training method according to an embodiment of the present invention;
fig. 3 is a flowchart of a processing method for an AI training task based on a block chain according to an embodiment of the present invention;
fig. 4 is a flowchart of a processing method for an AI training task based on a block chain according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an AI training task initiating node according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of an accounting node according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The terms "comprises" and "comprising," and any variations thereof, in the description and claims of this invention, are intended to cover non-exclusive inclusions, e.g., a process or an apparatus that comprises a list of steps is not necessarily limited to those structures or steps expressly listed but may include other steps or structures not expressly listed or inherent to such process or apparatus.
The artificial intelligence training method based on the block chain provided by the embodiment of the invention can be applied to the communication system shown in figure 1. As shown in fig. 1, the communication system includes: the node comprises an AI training task initiating node, an AI training node and a billing node, wherein the AI training task initiating node, the AI training node and the billing node are participating nodes in a block chain network. It is understood that the description is only illustrative and does not limit the number and types of nodes in the blockchain network. The accounting node can be one or a plurality of cloud servers, the cloud servers are a server cluster, a plurality of servers are arranged, the server cluster is similar to a universal computer framework, and the cloud servers comprise processors, hard disks, memories, system buses and the like. The AI training task initiating node or the AI training node may specifically be a user terminal, for example, a smartphone, a tablet computer, a personal computer, or the like. In addition, in embodiments of the present invention, the blockchain network is a decentralized, peer-to-peer (P2P) communication network.
The embodiment of the invention provides an artificial intelligence training method based on a block chain, and aims to solve the technical problems in the prior art.
The following describes the technical solutions of the present invention and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present invention will be described below with reference to the accompanying drawings.
Fig. 2 is a flowchart of an artificial intelligence training method based on a blockchain according to an embodiment of the present invention, and as shown in fig. 2, the method includes the following steps:
The AI training task in this embodiment may be specifically exemplified by a training task of an artificial intelligence model, and the training task needs to use a large amount of sample data as input to train the processing capability of the model, so that the accuracy of the model processing result meets the preset requirement. The model referred to in this embodiment may be embodied as any artificial intelligence model known in the art.
The AI training node in this embodiment provides a model training service, performs model training according to an AI training task initiated by the AI training task initiating node, and feeds back a training result to the AI training task initiating node.
And the accounting node is used for verifying the training result of the AI training node and recording the reward which the AI training node is entitled according to the training result reward scheme issued by the AI training task initiating node.
For example, in this embodiment, the AI training task initiating node may issue the training task to the blockchain network in the form of a task white paper. For example, in one embodiment, the AI training task initiating node monitors a user action, acquires an edited intelligent contract white paper from a user interaction interface after monitoring a task issuing action of the user, signs the intelligent contract white paper through a private key of the intelligent contract white paper, and broadcasts the signed intelligent contract white paper to the blockchain network, so that the AI training node in the blockchain network executes the training task. The intelligent contract white paper at least comprises a first training result detection sample for detecting a training result, or a link for storing the first training result detection sample, and a reward scheme of the training result, wherein the first training result detection sample only comprises the sample and does not comprise a correct calculation result corresponding to the sample. The rewarding scheme may include at least a lowest accuracy rate at which a reward can be obtained, a first correspondence relationship between an accuracy rate of a first training result satisfying the lowest accuracy rate and a reward amount, and a second correspondence relationship between a gradient of a subsequent training result and the reward amount, for example, in an embodiment, a training task may be completed first within a time limit, a training result with a model accuracy rate of 95% or more is rewarded with x tokens (tokens), a training result with a model accuracy rate of 90% or more is rewarded with y tokens, and for example, it may be specified that M tokens and the like are rewarded for each gradient of the subsequent training result. It is to be understood that this is by way of illustration and not by way of limitation.
After receiving an AI training task issued by an AI training task initiating node, an AI training node in the block chain starts its own computing cluster (for example, a Graphics Processing Unit (GPU) computing cluster) to load a training sample, or loads a training result detection sample according to a link issued by the AI training task initiating node, and performs model training to obtain model data. Or the existing model data meeting the requirements of the AI training task can be the existing model data, or the optimized model data can be obtained by optimizing the model data on the basis of the model data issued by other training nodes in the blockchain network. Further, after the model data is obtained, the AI training node calculates the first training result detection sample by using the model data to obtain a first calculation result corresponding to the first training result detection sample, signs the model data and the first calculation result by using a private key of the AI training node, and issues the data signed by the private key to the block chain network.
After the AI training task initiating node receives the model data and the first calculation result issued by the AI training node, issuing a second calculation result corresponding to the first training result detection sample in the blockchain network, wherein the calculation result is a correct result corresponding to the first training result detection sample.
And the accounting node in the block chain network normally records all records issued by each node in the block chain network and obtains a normal Token reward. Meanwhile, the accounting node receives an AI training task and a second calculation result issued by the AI training task initiating node in the blockchain network, and model data and a first calculation result issued by the AI training node in the blockchain network according to the AI training task. And comparing the first calculation result with the second calculation result to determine the accuracy of the model data. If the model data is the first training result meeting the lowest accuracy rate after the task is issued, determining the reward amount of the AI training node according to the first corresponding relation issued by the AI training task initiating node; and if the model data is a training result issued after the first training result, determining the reward amount of the AI training node based on the second corresponding relation issued by the AI training task initiating node. And after the reward amount of the AI training node is determined, recording the amount into the blockchain network. After that, if other AI training nodes issue new model data, the same method is followed for acceptance. And generates a corresponding reward record.
In this embodiment, once the reward amount is written into the blockchain network, the current AI training task initiating node or other AI training task initiating nodes may continue to issue the next training task and the corresponding reward amount, and the accounting node selects the model data issued by the AI training node that meets the condition according to the newly issued task and the reward scheme to perform reward.
In this embodiment, the AI training task initiating node not only rewards the most effective nodes, but sets the lowest accuracy rate at which the reward can be obtained, and the reward can be obtained every time the newly issued model data is promoted. The AI training nodes issue model data each time, the accounting nodes determine whether to reward according to the accuracy of the model data, and other AI training nodes can continue training and optimizing on the basis of the issued model data. And the AI training task initiating node sets a training target and an award amount of the next stage. Or other AI training initiating nodes relay the set award amount. Thus, all AI training nodes are trained on the basis of other training nodes, and can be rewarded as long as the AI training nodes can be promoted.
In this embodiment, the AI training task initiating node issues an AI training task carrying a first training result detection sample and a training result reward scheme into the blockchain network, and after receiving model data issued by the AI training task by the AI training node based on the AI training task and a first calculation result of the model data for the first training result detection sample, issues a second calculation result corresponding to the first training result detection sample in the blockchain network, so that the accounting node in the blockchain network can obtain the accuracy of the model data by comparing the first calculation result with the second calculation result, and record the reward amount to the AI training node according to the accuracy and the training result reward scheme, thereby not only realizing AI training task processing based on the blockchain network, but also reducing the calculation amount of the accounting node in the model accuracy verification process, the computational pressure of the accounting node is reduced.
Fig. 3 is a flowchart of a processing method for an AI training task based on a block chain according to an embodiment of the present invention, and as shown in fig. 3, on the basis of the above embodiment, the method includes:
301, an AI training task initiating node issues an AI training task in a blockchain network, where the AI training task includes a first training result detection sample and a training result reward scheme, where the first training result detection sample does not include a calculation result corresponding to a sample.
And 303, the AI training task initiating node issues a second calculation result corresponding to the first training result detection sample and a second training result detection sample for verifying the accuracy of the model data in the blockchain network, so that the accounting node in the blockchain network obtains the accuracy of the model data by comparing the first calculation result with the second calculation result, verifies the accuracy by using the second training result detection sample, and records the reward amount to the AI training node according to the training result reward scheme and the accuracy when the verification passes.
In this embodiment, the AI training task initiating node issues a first training result detection sample, the first training result detection sample does not include a calculation result of the sample, the AI training node calculates a first calculation result corresponding to the first training result detection sample according to model data obtained by training of the AI training node itself, and issues the first calculation result to the blockchain network. Once the AI training node publishes the model data and the first calculation result in the blockchain network, the AI training task initiating node publishes a second calculation result of the first training result detection sample in the blockchain network, comparison is carried out by the accounting node, and the accuracy is obtained according to the comparison result. Thus, the accuracy of the model data is calculated according to the model data and the sample without using a billing node. The calculation amount of the accounting node is saved. Meanwhile, the accounting node can randomly select a plurality of samples from the second training result detection samples issued by the AI training task initiating node to verify the accuracy of the model data, so that the first calculation result is generated by the model data instead of being obtained by cheating in a manual mode or other modes, and the reliability of the training result is improved.
Fig. 4 is a flowchart of a processing method for an AI training task based on a block chain according to an embodiment of the present invention, and as shown in fig. 4, on the basis of the above embodiment, the method includes:
And step 403, the accounting node receives a second calculation result of the first training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result.
And step 404, the accounting node compares the first calculation result with the second calculation result to obtain the accuracy of the model data, and records the reward amount to the AI training node according to the training result reward scheme and the accuracy.
Optionally, the method may further include:
and the accounting node receives a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
Optionally, after the accounting node receives a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, the method further includes:
and the accounting node verifies the accuracy of the model data based on the second training result detection sample.
The implementation manner and the beneficial effects of this embodiment are similar to those of the above method embodiment, and are not described herein again.
Fig. 5 is a schematic structural diagram of an AI training task initiating node according to an embodiment of the present invention, and as shown in fig. 5, a node 50 includes: a processor 51 and a memory 52, the memory 52 having instructions stored therein, which when executed by the processor 51, perform the following:
issuing an AI training task in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result reward scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample; receiving model data issued by an AI training node based on the AI training task and a first calculation result of the model data aiming at the first training result detection sample; and issuing a second calculation result corresponding to the first training result detection sample in the blockchain network so that the accounting node in the blockchain network obtains the accuracy of the model data by comparing the first calculation result with the second calculation result, and recording the reward amount of the AI training node according to the training result reward scheme and the accuracy.
Optionally, the processor is further configured to:
and issuing a second training result detection sample in the blockchain network so that the accounting node verifies the accuracy of the model data based on the second training result detection sample, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
The node provided in this embodiment can execute the method in the embodiment of fig. 2 or fig. 3, and the execution manner and the beneficial effect are similar, which are not described herein again.
Fig. 6 is a schematic structural diagram of an accounting node according to an embodiment of the present invention, and as shown in fig. 6, the node 60 includes: a processor 61 and a memory 62, the memory 62 having stored therein instructions that when executed by the processor 61 perform the following:
receiving an AI training task issued by an AI training task initiating node in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample; receiving model data issued by an AI training node according to the AI training task and a first calculation result of the model data aiming at the first training result detection sample; receiving a second calculation result of the first training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result; and comparing the first calculation result with the second calculation result to obtain the accuracy of the model data, and recording the reward amount of the AI training node according to the training result reward scheme and the accuracy.
Optionally, the processor is further configured to:
and receiving a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
The node provided in this embodiment can execute the method in the embodiment of fig. 4, and the execution manner and the beneficial effect are similar, which are not described herein again.
Embodiments of the present invention also provide a computer-readable storage medium, which includes instructions that, when executed on a computer, cause the computer to perform the method provided in any of the above embodiments.
Finally, it should be noted that, as one of ordinary skill in the art will appreciate, all or part of the processes of the methods of the embodiments described above may be implemented by hardware related to instructions of a computer program, where the computer program may be stored in a computer-readable storage medium, and when executed, the computer program may include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), a Random Access Memory (RAM), or the like.
Each functional unit in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a separate product, may also be stored in a computer readable storage medium. The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
The above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A processing method of an AI training task based on a block chain is characterized by comprising the following steps:
an AI training task initiating node issues an AI training task in a blockchain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample;
the AI training task initiating node receives model data issued by the AI training node based on the AI training task and a first calculation result of the model data aiming at the first training result detection sample;
and the AI training task initiating node issues a second calculation result corresponding to the first training result detection sample in the blockchain network so that the accounting node in the blockchain network obtains the accuracy of the model data by comparing the first calculation result with the second calculation result, and records the reward amount of the AI training node according to the training result reward scheme and the accuracy.
2. The method of claim 1, wherein the AI training task initiating node receives model data issued by an AI training node based on the AI training task, and after the model data detects a first computation result of a sample for the first training result, the method further comprises:
and the AI training task initiating node issues a second training result detection sample in the blockchain network so that the accounting node verifies the accuracy of the model data based on the second training result detection sample, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
3. A processing method of an AI training task based on a block chain is characterized by comprising the following steps:
the method comprises the steps that an accounting node receives an AI training task issued by an AI training task initiating node in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to a sample;
the accounting node receives model data issued by an AI training node according to the AI training task and a first calculation result of the model data aiming at the first training result detection sample;
The accounting node receives a second calculation result of the first training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result;
and the accounting node compares the first calculation result with the second calculation result to obtain the accuracy of the model data, and records the reward amount of the AI training node according to the training result reward scheme and the accuracy.
4. The method of claim 3, further comprising:
and the accounting node receives a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
5. The method of claim 4, wherein the accounting node receives a second training result test sample issued by the AI training task initiating node after receiving the model data and the first computation result, the method further comprising:
and the accounting node verifies the accuracy of the model data based on the second training result detection sample.
6. An AI training task initiating node, comprising: a processor and a memory, the memory having instructions stored therein that when executed by the processor perform the following:
issuing an AI training task in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result reward scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample;
receiving model data issued by an AI training node based on the AI training task and a first calculation result of the model data aiming at the first training result detection sample;
and issuing a second calculation result corresponding to the first training result detection sample in the blockchain network so that the accounting node in the blockchain network obtains the accuracy of the model data by comparing the first calculation result with the second calculation result, and recording the reward amount of the AI training node according to the training result reward scheme and the accuracy.
7. The AI training task initiating node of claim 6, wherein the processor is further configured to:
And issuing a second training result detection sample in the blockchain network so that the accounting node verifies the accuracy of the model data based on the second training result detection sample, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
8. An accounting node, comprising: a processor and a memory, the memory having instructions stored therein that when executed by the processor perform the following:
receiving an AI training task issued by an AI training task initiating node in a block chain network, wherein the AI training task comprises a first training result detection sample and a training result rewarding scheme, and the first training result detection sample does not contain a calculation result corresponding to the sample;
receiving model data issued by an AI training node according to the AI training task and a first calculation result of the model data aiming at the first training result detection sample;
receiving a second calculation result of the first training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result;
and comparing the first calculation result with the second calculation result to obtain the accuracy of the model data, and recording the reward amount of the AI training node according to the training result reward scheme and the accuracy.
9. The accounting node of claim 8, wherein the processor is further configured to:
and receiving a second training result detection sample issued by the AI training task initiating node after receiving the model data and the first calculation result, wherein the second training result detection sample comprises a sample and a corresponding calculation result.
10. A computer-readable storage medium comprising instructions that, when executed on a computer, cause the computer to perform the method of any of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010624145.4A CN111858756A (en) | 2020-06-30 | 2020-06-30 | Processing method, node and medium for AI training task based on block chain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010624145.4A CN111858756A (en) | 2020-06-30 | 2020-06-30 | Processing method, node and medium for AI training task based on block chain |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111858756A true CN111858756A (en) | 2020-10-30 |
Family
ID=72988975
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010624145.4A Pending CN111858756A (en) | 2020-06-30 | 2020-06-30 | Processing method, node and medium for AI training task based on block chain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111858756A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107864198A (en) * | 2017-11-07 | 2018-03-30 | 济南浪潮高新科技投资发展有限公司 | A kind of block chain common recognition method based on deep learning training mission |
| CN109918444A (en) * | 2019-02-01 | 2019-06-21 | 上海尚阵智能科技有限公司 | Training/verifying/management method/system, medium and equipment of model result |
| CN110675361A (en) * | 2019-08-16 | 2020-01-10 | 北京百度网讯科技有限公司 | Method and device for establishing video detection model and video detection |
| CN111125784A (en) * | 2019-12-24 | 2020-05-08 | 山东爱城市网信息技术有限公司 | Artificial intelligence training model method, device and medium based on block chain |
| US20210142908A1 (en) * | 2018-06-22 | 2021-05-13 | H-Labs Gmbh | A method to determine a degree of abnormality, a respective computer readable medium and a distributed cancer analysis system |
-
2020
- 2020-06-30 CN CN202010624145.4A patent/CN111858756A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107864198A (en) * | 2017-11-07 | 2018-03-30 | 济南浪潮高新科技投资发展有限公司 | A kind of block chain common recognition method based on deep learning training mission |
| US20210142908A1 (en) * | 2018-06-22 | 2021-05-13 | H-Labs Gmbh | A method to determine a degree of abnormality, a respective computer readable medium and a distributed cancer analysis system |
| CN109918444A (en) * | 2019-02-01 | 2019-06-21 | 上海尚阵智能科技有限公司 | Training/verifying/management method/system, medium and equipment of model result |
| CN110675361A (en) * | 2019-08-16 | 2020-01-10 | 北京百度网讯科技有限公司 | Method and device for establishing video detection model and video detection |
| CN111125784A (en) * | 2019-12-24 | 2020-05-08 | 山东爱城市网信息技术有限公司 | Artificial intelligence training model method, device and medium based on block chain |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110618924B (en) | Link pressure testing method of web application system | |
| CN108648000B (en) | Method and device for evaluating user retention life cycle and electronic equipment | |
| CN111949394A (en) | Method, system and storage medium for sharing computing power resource | |
| CN108665363B (en) | Block chain consensus achieving device | |
| CN109739527B (en) | Method, device, server and storage medium for client gray scale release | |
| CN105357167A (en) | Service processing method and device | |
| CN112418259A (en) | Method for configuring real-time rules based on user behaviors in live broadcast process, computer equipment and readable storage medium | |
| CN111158887A (en) | Centralized data distributed processing method and device | |
| CN112291119A (en) | Block chain network testing method, device, medium and electronic equipment | |
| CN110743169B (en) | Anti-cheating method and system based on block chain | |
| CN111858752A (en) | Artificial intelligence training method and device based on block chain and storage medium | |
| CN111858755A (en) | Processing method, node and medium for AI training task based on block chain | |
| CN113486118A (en) | Consensus node selection method and device | |
| CN111340574B (en) | Risk user identification method and device and electronic equipment | |
| CN104135525B (en) | The resource expansion method and apparatus of cloud platform ELB components | |
| CN112699049A (en) | Block chain network testing method, device, medium and electronic equipment | |
| CN111858756A (en) | Processing method, node and medium for AI training task based on block chain | |
| CN111858753A (en) | Block chain-based training parameter processing method, device and storage medium | |
| CN113821443A (en) | Application program function detection method, device, equipment and storage medium | |
| CN113673811A (en) | Session-based online learning performance evaluation method and device | |
| CN112654077B (en) | Energy-saving method and device, and computer storage medium | |
| CN114242181A (en) | Desert sand concrete strength prediction model training method, device, equipment and medium | |
| CN113742187A (en) | Capacity prediction method, device, equipment and storage medium of application system | |
| CN113342795B (en) | Data checking method and device in application program, electronic equipment and storage medium | |
| CN111711537B (en) | Method, device and equipment for updating standby main node list |
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 |