CN111723147B - Block chain-based data training method, device and equipment and storage medium - Google Patents
Block chain-based data training method, device and equipment and storage medium Download PDFInfo
- Publication number
- CN111723147B CN111723147B CN201910217423.1A CN201910217423A CN111723147B CN 111723147 B CN111723147 B CN 111723147B CN 201910217423 A CN201910217423 A CN 201910217423A CN 111723147 B CN111723147 B CN 111723147B
- Authority
- CN
- China
- Prior art keywords
- target
- model
- blockchain
- data
- model training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides a data training method, device and equipment based on a blockchain, and a storage medium, wherein the method comprises the following steps: receiving a model training task sent by a client; training a target model according to the model training task; target data associated with the target model is stored to a specified block in a specified blockchain according to a blockchain protocol of the specified blockchain. The client can search the target data related to the target model on the appointed blockchain, and the processing data quantity of the server is reduced.
Description
Technical Field
The present disclosure relates to the field of computer technologies, and in particular, to a data training method, apparatus and device based on a blockchain, and a storage medium.
Background
With the development of scientific technology, neural network models are excellent in tasks such as classification, detection and recognition, and a platform for training the models is generated.
In the related data training mode, after a user initiates a training task to a server serving as a training platform side through a client, the server executes model training operation according to the training task, and then stores a model locally and sends the model to the user when needed.
In the above manner, when the client downloads or views the trained model or other related data, the client needs to search through the server, so as to increase the processing data volume of the server.
Disclosure of Invention
In view of the above, the present application provides a data training method, device and equipment based on a blockchain, and a storage medium, where a client can search target data related to a target model on a specified blockchain, so as to reduce the processing data volume of a server.
The first aspect of the present application provides a data training method based on a blockchain, the method comprising:
receiving a model training task sent by a client;
training a target model according to the model training task;
target data associated with the target model is stored to a specified block in a specified blockchain according to a blockchain protocol of the specified blockchain.
According to one embodiment of the present application, the method further comprises:
and sending the storage position of the target data to the client so that the client searches the target data in the blockchain according to the storage position.
According to one embodiment of the application, the model training task carries a user identifier;
storing target data related to the target model to a specified block of a specified blockchain according to a blockchain protocol of the specified blockchain includes:
selecting a target blockchain protocol corresponding to the user identification from a plurality of blockchain protocols of the designated blockchain;
storing target data related to the target model to a specified block in the specified blockchain according to the target blockchain protocol.
According to one embodiment of the application, the target data comprises at least a target model and/or parameter information of the target model, wherein the parameter information comprises at least: the length of time used to train the target model and/or the performance parameters of the target model.
According to one embodiment of the present application, before the receiving the model training task sent from the client, the method further includes: receiving user identification and model training data sent by the client, and recording the user identification and corresponding first operation information, wherein the first operation information is used for indicating that the model training data is sent by the client;
the model training task carries a user identifier; storing target data related to the target model to a specified block of a specified blockchain according to a blockchain protocol of the specified blockchain, comprising:
generating second operation information corresponding to a user identifier carried by the model training task, and acquiring first operation information corresponding to the user identifier carried by the model training task, wherein the second operation information is used for indicating that the client side sends the model training task;
determining target data at least comprising the second operation information and the acquired first operation information;
the target data is stored to a specified block of a specified blockchain according to a blockchain protocol of the specified blockchain.
According to one embodiment of the present application, after receiving the user identification and the model training data sent by the client, the method further includes: storing the model training data corresponding to the user identification;
training a target model according to the model training task comprises:
obtaining locally recorded model training data corresponding to the user identification according to the user identification carried by the model training task;
and training the target model by using the acquired model training data.
A second aspect of the present application provides a blockchain-based data training device, the device comprising:
the task receiving module is used for receiving a model training task sent by the client;
the model training module is used for training a target model according to the model training task;
and the data storage module is used for storing the target data related to the target model to the designated block of the designated block chain according to the block chain protocol of the designated block chain.
According to one embodiment of the present application, the apparatus further comprises:
and the storage position sending module is used for sending the storage position of the target data to the client so that the client searches the target data in the blockchain according to the storage position.
According to one embodiment of the application, the model training task carries a user identifier;
the data storage module includes:
a blockchain protocol selection unit, configured to select a target blockchain protocol corresponding to the user identifier from a plurality of blockchain protocols of the specified blockchain;
and the data storage unit is used for storing target data related to the target model to a specified block in the specified block chain according to the target block chain protocol.
According to one embodiment of the application, the target data comprises at least a target model and/or parameter information of the target model, wherein the parameter information comprises at least: the length of time used to train the target model and/or the performance parameters of the target model.
According to one embodiment of the present application, before the task receiving module, the apparatus further includes: the receiving and recording module is used for receiving the user identification and the model training data sent by the client, and recording the user identification and corresponding first operation information, wherein the first operation information is used for indicating that the model training data is sent by the client;
the model training task carries a user identifier; the data storage module includes:
an operation information determining unit, configured to generate second operation information corresponding to a user identifier carried by the model training task, and obtain first operation information corresponding to the user identifier carried by the model training task, where the second operation information is used to indicate that the client side has sent the model training task;
a target data determining unit that determines target data including at least the second operation information and the acquired first operation information;
and the target data storage unit is used for storing the target data to the designated block in the designated block chain according to the block chain protocol of the designated block chain.
According to one embodiment of the present application, the receiving and recording module is further configured to, after receiving the user identification and the model training data sent from the client,: storing the model training data corresponding to the user identification;
the model training module comprises:
the model training data acquisition unit is used for acquiring locally recorded model training data corresponding to the user identification according to the user identification carried by the model training task;
and the target model training unit is used for training the target model by using the acquired model training data.
A third aspect of the present application provides an electronic device, comprising a processor and a memory; the memory stores a program that can be called by the processor; wherein, when the processor executes the program, the blockchain-based data training method according to the foregoing embodiment is implemented.
A fourth aspect of the present application provides a machine-readable storage medium having stored thereon a program which, when executed by a processor, implements a blockchain-based data training method as described in the previous embodiments.
The embodiment of the application has the following beneficial effects:
in the embodiment of the application, after training the target model according to the model training task initiated by the client, the server can store the target data related to the target model on the designated block of the designated block chain according to the block chain protocol deployed on the designated block chain, when the client needs to check or download the target data, the client only needs to check on the designated block chain, the server serving as a training platform side is not required to check, the processing data amount of the server is reduced, and in addition, because the target data is recorded on the designated block chain, the target data can be prevented from being manually modified, and the consistent target data can be ensured whenever the target data is checked or downloaded.
Drawings
FIG. 1 is a flow chart of a blockchain-based data training method in accordance with an embodiment of the present application;
FIG. 2 is a schematic diagram of interactions in a blockchain-based data training method in accordance with an embodiment of the present application;
FIG. 3 is a block diagram of a block chain based data training device according to one embodiment of the present application;
fig. 4 is a block diagram of an electronic device according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples are not representative of all implementations consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present application as detailed in the accompanying claims.
The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in this application and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any or all possible combinations of one or more of the associated listed items.
It should be understood that although the terms first, second, third, etc. may be used herein to describe various devices, these information should not be limited by these terms. These terms are only used to distinguish one device from another of the same type. For example, a first device could also be termed a second device, and, similarly, a second device could also be termed a first device, without departing from the scope of the present application. The word "if" as used herein may be interpreted as "at … …" or "at … …" or "responsive to a determination", depending on the context.
In order to make the description of the present application clearer and concise, some technical terms in the present application are explained below:
neural network: a technique for simulating the abstraction of brain structure features that a network system is formed by complex connection of a great number of simple functions, which can fit extremely complex functional relation, and generally includes convolution/deconvolution operation, activation operation, pooling operation, addition, subtraction, multiplication and division, channel merging and element rearrangement. Training the network with specific input data and output data, adjusting the connections therein, and allowing the neural network to learn the mapping between the fitting inputs and outputs.
Deep learning: a neural network learning technology learns relevant knowledge from data for subsequent prediction by simulating learning behaviors of human brain.
Reinforcement learning, which is a machine learning method, is essentially a mapping learning from environmental states to behavioral actions so that the agent takes optimal decisions based on maximum returns.
Deep reinforcement learning (Deep Reinforcement Learning): is a brand new algorithm combining deep learning and reinforcement learning, thereby realizing end-to-end learning from environmental state to behavior action. In short, after the neural network is subjected to deep reinforcement learning, the state of receiving input can be directly output for action.
Blockchain: a special distributed database has the characteristics of decentralization, openness, autonomy, non-tamperability of information and anonymity, and can record continuously-growing and non-tamperable data by maintaining a chain structure of blocks. Blockchain technology is a technology that collectively maintains a reliable database by means of de-centralization and de-trust.
Public chain (Public Blockchain): the public blockchain refers to a public blockchain which is read, sent and effectively confirmed by anyone in the world, can anonymously access a network without registration and authorization, and has the characteristics of decentralization, neutrality, openness, non-falsification and the like, such as a bit coin block chain, an Ethernet intelligent contract and the like, which are all intelligent public chains.
Open platform: refers to a software system disclosing its Application Programming Interface (API) or function (function) through which an external program may add to the functionality of the software system or use the software system resources.
The data training method of the present application is based on blockchain. In the data training method, the appointed blockchain is used for storing target data related to the target model obtained through training, so that the client can search the target data on the appointed blockchain later, a server serving as a training platform side is not required to search the target data, and the data processing capacity of the server is reduced.
The following describes the blockchain-based data training method in the embodiment of the present application in more detail, but should not be limited thereto.
In one embodiment, referring to fig. 1 and 2, a blockchain-based data training method includes the steps of:
s100: receiving a model training task sent by a client;
s200: training a target model according to the model training task;
s300: target data associated with the target model is stored to a specified block in a specified blockchain according to a blockchain protocol of the specified blockchain.
The data training method based on the blockchain can be applied to a server, and a blockchain application platform can be loaded on the server and used as a training platform side. The server may be one computer device or a plurality of computer devices, and is not particularly limited.
The blockchain application platform is an application platform which can realize related functions according to a specified blockchain and can be an open platform. The specified blockchain may be a public chain, a private chain, or a federated chain. In this embodiment, the specified blockchain may be selected as a public chain, such as ethernet.
In step S100, a model training task transmitted from a client is received.
The client may first establish a connection with the server before sending the model training task to the server. The manner of establishing connection between the client and the server is not limited as long as communication between the client and the server can be achieved. Of course, the server may be connected to multiple clients to provide model training services for the multiple clients.
The server receives model training tasks from the client. The model training task may be used to indicate that the model to be trained needs to be trained according to the training mode. The training regimen and/or model to be trained may be specified by a model training task. When only one model to be trained exists in the server and a plurality of models to be trained exist in the server, the model training task can only specify a training mode; alternatively, the model training task may specify only the model to be trained when there are only a plurality of models to be trained and only one model to be trained on the server. As may be desired.
In step S200, a target model is trained according to the model training task.
Models to be trained can be pre-stored on the server, and the types can be one or more than two. The models to be trained are untrained neural networks, and the architectures of different neural networks can be different, so that the neural networks after training can realize functional diversification, such as target detection, character recognition and the like. The layer structure and the number of layers of the neural network are not limited.
The server can be preset with a plurality of training modes, namely the learning mode of the neural network, which can comprise deep learning, reinforcement learning, deep reinforcement learning and other modes, and can be further subdivided into a plurality of learning modes, such as a first deep learning mode, a second deep learning mode and the like.
Of course, the model to be trained and the training mode can also be set on the external device, and the model to be trained and the training mode can be acquired from the external device when the model training task is received.
The model training task may specify a certain model to be trained and a certain training mode, for example, a first model to be trained and a first training mode, and then the server may acquire the first model to be trained and train the first model to be trained in the first training mode to obtain the target model. The architecture of the target model obtained through training is the same as that of the first model to be trained, the first model to be trained is trained so that network parameters of the first model to be trained are optimized to obtain the target model, and therefore the target model can achieve corresponding functions.
In step S300, target data associated with the target model is stored to a specified block in a specified blockchain according to a blockchain protocol of the specified blockchain.
The server may implement step S300 according to a blockchain protocol deployed on the specified blockchain. The blockchain protocol may be presented in the form of a smart contract code that is executed independently on each node in the specified blockchain in a prescribed manner, invoking the smart contract code to implement the corresponding function of the blockchain protocol. The specified block on the specified block chain can be determined according to the block chain protocol, and then the target data is recorded on the specified block.
The target data is data related to the target model, and may include the target model, parameters embodying the current training situation, and the like. The target data may be some data required by the user, and after storing the target data in a specified block in the specified blockchain, the user may search the specified block in the specified blockchain as required to view or download the target data or some data therein.
In the embodiment of the application, after training the target model according to the model training task initiated by the client, the server can store the target data related to the target model on the designated block of the designated block chain according to the block chain protocol deployed on the designated block chain, when the client needs to check or download the target data, the client only needs to check on the designated block chain, the server serving as a training platform side is not required to check, the processing data amount of the server is reduced, and in addition, because the target data is recorded on the designated block chain, the target data can be prevented from being manually modified, and the consistent target data can be ensured whenever the target data is checked or downloaded.
In one embodiment, the above method flow may be performed by the blockchain-based data training device 100, and as shown in fig. 3, the blockchain-based data training device 100 mainly includes 3 modules: a task receiving module 101, a model training module 102 and a data storage module 103. The task receiving module 101 is configured to perform the step S100, the model training module 102 is configured to perform the step S200, and the data storage module 103 is configured to perform the step S300.
In one embodiment, referring to fig. 2, the method further comprises the steps of:
s400: and sending the storage position of the target data to the client so that the client searches the target data in the blockchain according to the storage position.
After storing the target data on the specified block of the specified blockchain according to the blockchain protocol, the above step S400 may also be implemented according to the blockchain protocol, that is, the storage location of the target data is sent to the client according to the blockchain protocol.
The storage location of the target data, i.e., the location of the target data in the specified blockchain, may be a block identification of the specified block in the specified blockchain in which the target data is stored. After the storage location of the target data is sent to the client, the client may find the target data on the specified blockchain according to the storage location, for further viewing or downloading, and so on.
When the client searches data, the storage position of the data to be searched can be sent to the blockchain protocol of the appointed blockchain, so that the block corresponding to the storage position can be searched on the appointed blockchain according to the blockchain protocol, and the data on the block can be acquired and returned to the client.
The functionality of the blockchain protocol may be as desired. In this embodiment, the blockchain protocol may implement storing the target data on a specified block of a specified blockchain and sending the storage location of the target data to the client. Of course, the blockchain protocol may also implement sending the storage location to the server, and then the storage location is sent to the client by the server, which is not particularly limited. The blockchain protocol may also enable other information to be sent to clients, or servers.
The blockchain protocol may be created by a server where the application platform resides. Taking the ethernet as an example, to create a blockchain protocol in the ethernet, the blockchain protocol needs to be written, changed into byte codes, deployed to a blockchain, and the like. Invoking the blockchain protocol from the ethernet is initiating a transaction directed to the blockchain protocol address.
The blockchain protocol may be defined in the form of code. Taking the ethernet as an example, some complex logic can be created and called in the ethernet network, the ethernet is used as a core of a programmable blockchain as an ethernet virtual machine, and each ethernet node can run the virtual machine. Issuing and invoking blockchain protocols in an ethernet network is performed on virtual machines, with blockchain protocols running in a distributed manner in virtual machines of each node in the ethernet network. For example, in an alternative embodiment where the virtual machine runs directly on virtual machine code (virtual machine bytecode), the blockchain protocol deployed on the blockchain may be in the form of virtual machine bytecode.
In one embodiment, the model training task carries a user identification;
in step S300, storing target data related to the target model in a specified blockchain according to a blockchain protocol of the specified blockchain includes the steps of:
s301: selecting a target blockchain protocol corresponding to the user identification from a plurality of blockchain protocols of the designated blockchain;
s302: storing target data related to the target model to a specified block in the specified blockchain according to the target blockchain protocol.
The user identification is an identification of the current user on the client, indicating that the model training task was sent by the user to the server via the client.
Each blockchain protocol is created on a specified blockchain before receiving a model training task, or on a specified blockchain after receiving the training task, and the specific creation time is not limited.
In step S301, a blockchain protocol corresponding to the user identifier carried in the model training task may be selected from the multiple blockchain protocols of the specified blockchain as the target blockchain protocol. In step S302, target data associated with a target model is stored to a specified block in the specified blockchain according to the target blockchain protocol.
One user identification may correspond to one blockchain protocol, each blockchain protocol corresponding to a respective user identification, which may prevent users from invoking other blockchain protocols by mistake, although this is not limiting.
In an embodiment, the target data comprises at least a target model, and/or parameter information of the target model, wherein the parameter information comprises at least: the length of time used to train the target model, and/or the performance parameters of the target model.
The time length used for training the target model can be obtained together when the target model is trained. The performance parameters of the target model include, for example: the accuracy, the robustness and the like of the target model can be obtained when the target model is trained, or can be obtained after further performance verification is carried out on the target model.
The time length and the performance parameters can embody the condition of the model training task, and after the condition is stored on the appointed block of the appointed block chain, the client can search on the appointed block chain, and can be used as a reference for evaluating the model training task by a user.
Of course, the duration and the performance parameter are related to the cost of the transaction generated by the model training task, for example, the longer the duration, the better the performance parameter indicates performance, the more expensive the transaction. After the client searches the target data in the blockchain, the time length and the performance parameters can be checked and used as references of transaction fees.
The target model, the time length and the performance parameters are stored on the appointed block of the appointed blockchain, and due to the non-tamperable characteristic of the blockchain, the target model, the time length and the performance parameters can be ensured to be always corresponding, and the target model, the time length and the performance parameters which are searched by a user each time are the same.
In one embodiment, prior to step S100, the method further comprises step S010: receiving user identification and model training data sent by the client, and recording the user identification and corresponding first operation information, wherein the first operation information is used for indicating that the model training data is sent by the client;
the model training task carries a user identifier; in step S300, storing target data related to the target model to a specified block in a specified blockchain according to a blockchain protocol of the specified blockchain, including:
generating second operation information corresponding to a user identifier carried by the model training task, and acquiring first operation information corresponding to the user identifier carried by the model training task, wherein the second operation information is used for indicating that the client side sends the model training task;
determining target data at least comprising the second operation information and the acquired first operation information;
the target data is stored to a specified block of a specified blockchain according to a blockchain protocol of the specified blockchain.
In step S010, since the model training data required for training the model is generally large in data size, the time required for transmitting the model training data to the server is relatively long, and the model training data is uploaded to the server in advance by the client, so that the server can be ensured to be executed in time when receiving the model training task, and the problem of too slow response caused by low data transmission rate of the model training task is avoided.
After model training data is received, first operation information corresponding to the user identification sent by the client can be generated, and the first operation information is recorded corresponding to the received user identification, so that when a model training task is received, the first operation information corresponding to the user identification carried by the model training task can be searched on a server.
In step S300, the model training task has been received, the model training task carries a user identifier, and the second operation information corresponding to the user identifier carried by the model training task may be generated, and the first operation information corresponding to the user identifier carried by the model training task may be obtained. The second operation information may also be recorded in the server in correspondence with the user identification.
And determining target data at least comprising the second operation information and the acquired first operation information. Of course, the object data may also include other information, such as the object model and/or parameter information of the object model as described in the previous embodiments. The target data is stored to a specified block of a specified blockchain according to a blockchain protocol of the specified blockchain.
The first operation information is used for the client to send the model training data, namely, the first operation information is an operation record about uploading the model training data by a user; the second operation information is used for indicating that the client side sends the model training data, namely the second operation information is an operation record about a user sending a model training task. The first operation information and the second operation information are stored as a part of target data to a designated block of a designated block chain, so that a user can conveniently inquire about related operation records of the user on a training platform side.
In one embodiment, after step S010, the method further comprises step S011: storing the model training data corresponding to the user identification;
in step S200, training the target model according to the model training task includes the following steps:
s201: obtaining locally recorded model training data corresponding to the user identification according to the user identification carried by the model training task;
s202: and training the target model by using the acquired model training data.
Because the model training data is stored corresponding to the user identifier, the model training data corresponding to the user identifier can be obtained according to the user identifier carried by the model training task to train the model.
The specific way of training the target model by using the obtained model training data is not limited, and training can be performed according to a preset training way. For example, the model training data may include an input sample and an output sample, where the input sample is used as an input of the model to be trained, and the corresponding output sample is used as an output of the model to be trained, so as to optimize network parameters of the model to be trained, and obtain the target model.
The present application also provides a blockchain-based data training device, referring to fig. 3, the blockchain-based data training device 100 includes:
a task receiving module 101, configured to receive a model training task sent from a client;
the model training module 102 is configured to train a target model according to the model training task;
and the data storage module 103 is used for storing the target data related to the target model to the designated block in the designated block chain according to the block chain protocol of the designated block chain.
In one embodiment, the apparatus further comprises:
and the storage position sending module is used for sending the storage position of the target data to the client so that the client searches the target data in the blockchain according to the storage position.
In one embodiment, the model training task carries a user identification;
the data storage module includes:
a blockchain protocol selection unit, configured to select a target blockchain protocol corresponding to the user identifier from a plurality of blockchain protocols of the specified blockchain;
and the data storage unit is used for storing target data related to the target model to a specified block in the specified block chain according to the target block chain protocol.
In an embodiment, the target data comprises at least a target model and/or parameter information of the target model, wherein the parameter information comprises at least: the length of time used to train the target model and/or the performance parameters of the target model.
In one embodiment, before the task receiving module, the apparatus further includes: the receiving and recording module is used for receiving the user identification and the model training data sent by the client, and recording the user identification and corresponding first operation information, wherein the first operation information is used for indicating that the model training data is sent by the client;
the model training task carries a user identifier; the data storage module includes:
an operation information determining unit, configured to generate second operation information corresponding to a user identifier carried by the model training task, and obtain first operation information corresponding to the user identifier carried by the model training task, where the second operation information is used to indicate that the client side has sent the model training task;
a target data determining unit that determines target data including at least the second operation information and the acquired first operation information;
and the target data storage unit is used for storing the target data to the designated block in the designated block chain according to the block chain protocol of the designated block chain.
In one embodiment, the receiving and recording module is further configured to, after receiving the user identification and the model training data sent from the client,: storing the model training data corresponding to the user identification;
the model training module comprises:
the model training data acquisition unit is used for acquiring locally recorded model training data corresponding to the user identification according to the user identification carried by the model training task;
and the target model training unit is used for training the target model by using the acquired model training data.
The implementation process of the functions and roles of each unit in the above device is specifically shown in the implementation process of the corresponding steps in the above method, and will not be described herein again.
For the device embodiments, reference is made to the description of the method embodiments for the relevant points, since they essentially correspond to the method embodiments. The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements.
The application also provides electronic equipment, which comprises a processor and a memory; the memory stores a program that can be called by the processor; wherein the processor, when executing the program, implements a blockchain-based data training method as described in the foregoing embodiments.
Embodiments of the blockchain-based data training device may be applied to an electronic device. Taking software implementation as an example, the device in a logic sense is formed by reading corresponding computer program instructions in a nonvolatile memory into a memory by a processor of an electronic device where the device is located for operation. In terms of hardware, as shown in fig. 4, fig. 4 is a hardware structure diagram of an electronic device where the blockchain-based data training device 100 is located according to an exemplary embodiment of the present application, and in addition to the processor 510, the memory 530, the interface 520, and the nonvolatile storage 540 shown in fig. 4, the electronic device where the device 100 is located in the embodiment may generally include other hardware according to an actual function of the electronic device, which is not described herein again.
The present application also provides a machine-readable storage medium having stored thereon a program which, when executed by a processor, implements a blockchain-based data training method as in any of the previous embodiments.
The present application may take the form of a computer program product embodied on one or more storage media (including, but not limited to, magnetic disk storage, CD-ROM, optical storage, etc.) having program code embodied therein. Machine-readable storage media include both permanent and non-permanent, removable and non-removable media, and information storage may be implemented by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of machine-readable storage media include, but are not limited to: phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Disks (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, may be used to store information that may be accessed by the computing device.
The foregoing description of the preferred embodiments of the present invention is not intended to limit the invention to the precise form disclosed, and any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention are intended to be included within the scope of the present invention.
Claims (8)
1. A blockchain-based data training method, the method comprising:
receiving user identification and model training data sent by a client, and recording the user identification and corresponding first operation information, wherein the first operation information is used for indicating the client to send the model training data;
receiving a model training task sent by a client, wherein the model training task carries a user identifier;
training a target model according to the model training task;
generating second operation information corresponding to a user identifier carried by the model training task, and acquiring first operation information corresponding to the user identifier carried by the model training task, wherein the second operation information is used for indicating that the client side sends the model training task;
determining target data at least comprising the second operation information and the acquired first operation information;
selecting a target blockchain protocol corresponding to the user identification from a plurality of blockchain protocols of a designated blockchain;
storing the target data to a specified block in the specified blockchain according to the target blockchain protocol.
2. The blockchain-based data training method of claim 1, further comprising:
and sending the storage position of the target data to the client so that the client searches the target data in the blockchain according to the storage position.
3. The blockchain-based data training method of claim 1, wherein the target data includes at least a target model and/or parameter information of the target model, wherein the parameter information includes at least: the length of time used to train the target model and/or the performance parameters of the target model.
4. The blockchain-based data training method of claim 1, wherein after receiving the user identification, model training data from the client, the method further comprises: storing the model training data corresponding to the user identification;
training a target model according to the model training task comprises:
obtaining locally recorded model training data corresponding to the user identification according to the user identification carried by the model training task;
and training the target model by using the acquired model training data.
5. A blockchain-based data training device, the device comprising:
the receiving and recording module is used for receiving the user identification and the model training data sent by the client, and recording the user identification and corresponding first operation information, wherein the first operation information is used for indicating that the model training data is sent by the client;
the task receiving module is used for receiving a model training task sent by the client, wherein the model training task carries a user identifier;
the model training module is used for training a target model according to the model training task;
an operation information determining unit, configured to generate second operation information corresponding to a user identifier carried by the model training task, and obtain first operation information corresponding to the user identifier carried by the model training task, where the second operation information is used to indicate that the client side has sent the model training task;
a target data determining unit that determines target data including at least the second operation information and the acquired first operation information;
a blockchain protocol selection unit, configured to select a target blockchain protocol corresponding to the user identifier from a plurality of blockchain protocols of a specified blockchain;
and the data storage unit is used for storing the target data to the designated block in the designated block chain according to the target block chain protocol.
6. The blockchain-based data training device of claim 5, further comprising:
and the storage position sending module is used for sending the storage position of the target data to the client so that the client searches the target data in the blockchain according to the storage position.
7. The blockchain-based data training device of claim 5, wherein the target data includes at least a target model and/or parameter information of the target model, wherein the parameter information includes at least: the length of time used to train the target model and/or the performance parameters of the target model.
8. The blockchain-based data training device of claim 7, wherein the receive record module, after receiving the user identification, model training data from the client transmission, is further to: storing the model training data corresponding to the user identification;
the model training module comprises:
the model training data acquisition unit is used for acquiring locally recorded model training data corresponding to the user identification according to the user identification carried by the model training task;
and the target model training unit is used for training the target model by using the acquired model training data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910217423.1A CN111723147B (en) | 2019-03-21 | 2019-03-21 | Block chain-based data training method, device and equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910217423.1A CN111723147B (en) | 2019-03-21 | 2019-03-21 | Block chain-based data training method, device and equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111723147A CN111723147A (en) | 2020-09-29 |
| CN111723147B true CN111723147B (en) | 2023-07-25 |
Family
ID=72562597
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910217423.1A Active CN111723147B (en) | 2019-03-21 | 2019-03-21 | Block chain-based data training method, device and equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111723147B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112668501A (en) * | 2020-12-30 | 2021-04-16 | 杭州趣链科技有限公司 | Automatic encoder feature extraction method and device based on block chain excitation |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107885762A (en) * | 2017-09-19 | 2018-04-06 | 北京百度网讯科技有限公司 | Intelligent big data system, the method and apparatus that intelligent big data service is provided |
| CN108323200A (en) * | 2018-01-25 | 2018-07-24 | 深圳前海达闼云端智能科技有限公司 | Data training method and device based on block chain, storage medium and block chain link points |
| US10095888B1 (en) * | 2018-06-04 | 2018-10-09 | Capital One Services, Llc | Secure decentralized system utilizing smart contracts, a blockchain, and/or a distributed file system |
| CN108780313A (en) * | 2016-01-20 | 2018-11-09 | 罗伯特·博世有限公司 | Increment related data digging technology for assembly line |
| CN108881355A (en) * | 2017-05-16 | 2018-11-23 | 纬创资通股份有限公司 | Monitoring method based on Internet of things architecture, fog operation terminal and Internet of things system |
| CN109032610A (en) * | 2017-06-08 | 2018-12-18 | 杭州海康威视数字技术股份有限公司 | A kind of program bag dispositions method, electronic equipment and distributed system |
| CN109413175A (en) * | 2018-10-19 | 2019-03-01 | 北京奇艺世纪科技有限公司 | A kind of information processing method, device and electronic equipment |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10984913B2 (en) * | 2012-04-27 | 2021-04-20 | Netspective Communications Llc | Blockchain system for natural language processing |
| US20180096175A1 (en) * | 2016-10-01 | 2018-04-05 | James L. Schmeling | Blockchain Enabled Packaging |
| US20170124497A1 (en) * | 2015-10-28 | 2017-05-04 | Fractal Industries, Inc. | System for automated capture and analysis of business information for reliable business venture outcome prediction |
| US10762479B2 (en) * | 2017-04-05 | 2020-09-01 | Samsung Sds Co., Ltd. | Method and system for processing blockchain-based real-time transaction |
| US10713963B2 (en) * | 2017-05-25 | 2020-07-14 | International Business Machines Corporation | Managing lifelong learner events on a blockchain |
| US11144828B2 (en) * | 2017-06-09 | 2021-10-12 | Htc Corporation | Training task optimization system, training task optimization method and non-transitory computer readable medium for operating the same |
| CN107291862A (en) * | 2017-06-12 | 2017-10-24 | 腾讯科技（深圳）有限公司 | Business datum storage method, device, storage medium and electronic equipment |
| US20190073701A1 (en) * | 2017-09-06 | 2019-03-07 | Tesloop, Inc. | Vehicle valuation model based on vehicle telemetry |
| CN108898219A (en) * | 2018-06-07 | 2018-11-27 | 广东工业大学 | A kind of neural network training method based on block chain, device and medium |
| CN109190769B (en) * | 2018-08-21 | 2021-12-24 | 深圳点宽网络科技有限公司 | Artificial intelligence training method based on block chain |
| CN109493216B (en) * | 2018-09-30 | 2021-02-09 | 北京小米移动软件有限公司 | Model training method, device, system and storage medium |
-
2019
- 2019-03-21 CN CN201910217423.1A patent/CN111723147B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108780313A (en) * | 2016-01-20 | 2018-11-09 | 罗伯特·博世有限公司 | Increment related data digging technology for assembly line |
| CN108881355A (en) * | 2017-05-16 | 2018-11-23 | 纬创资通股份有限公司 | Monitoring method based on Internet of things architecture, fog operation terminal and Internet of things system |
| CN109032610A (en) * | 2017-06-08 | 2018-12-18 | 杭州海康威视数字技术股份有限公司 | A kind of program bag dispositions method, electronic equipment and distributed system |
| CN107885762A (en) * | 2017-09-19 | 2018-04-06 | 北京百度网讯科技有限公司 | Intelligent big data system, the method and apparatus that intelligent big data service is provided |
| CN108323200A (en) * | 2018-01-25 | 2018-07-24 | 深圳前海达闼云端智能科技有限公司 | Data training method and device based on block chain, storage medium and block chain link points |
| US10095888B1 (en) * | 2018-06-04 | 2018-10-09 | Capital One Services, Llc | Secure decentralized system utilizing smart contracts, a blockchain, and/or a distributed file system |
| CN109413175A (en) * | 2018-10-19 | 2019-03-01 | 北京奇艺世纪科技有限公司 | A kind of information processing method, device and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111723147A (en) | 2020-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20190370659A1 (en) | Optimizing neural network architectures | |
| CN109344295B (en) | Distributed graph embedding method, device, equipment and system | |
| US20130185039A1 (en) | Monte-carlo planning using contextual information | |
| CN110168570A (en) | Machine learning virtual sensor process model for multiple sensors | |
| CN111325417B (en) | Method and device for realizing privacy protection and realizing multi-party collaborative updating of business prediction model | |
| CN104243590B (en) | Resource object recommends method and apparatus | |
| CN107832275A (en) | The generation method of intelligent Contract ID, apparatus and system in block chain | |
| JP7038790B2 (en) | Neural episode control | |
| CN110555148B (en) | User behavior evaluation method, computing device and storage medium | |
| CN107077507A (en) | A kind of information-pushing method, device and system | |
| US10915511B2 (en) | Processing path determination | |
| FR3044438A1 (en) | DECISION SUPPORT SYSTEM AND METHOD | |
| CN112925893A (en) | Conversational interest point recommendation method and device, electronic equipment and storage medium | |
| CN115774819B (en) | Point of interest recommendation method and system based on hierarchical cyclic neural network | |
| CN111723147B (en) | Block chain-based data training method, device and equipment and storage medium | |
| CN112187710A (en) | Method and device for sensing threat intelligence data, electronic device and storage medium | |
| CN109635257A (en) | A kind of configuration method and device of contract template | |
| CN113689270B (en) | Method for determining black product device, electronic device, storage medium, and program product | |
| CN107517139A (en) | Data processing method and system, device test system and bus test system | |
| CN107231398A (en) | Information interacting method and device | |
| CN113726545B (en) | Network traffic generation method and device for generating countermeasure network based on knowledge enhancement | |
| EP3446258A1 (en) | Model-free control for reinforcement learning agents | |
| CN111507541B (en) | Goods quantity prediction model construction method, goods quantity measurement device and electronic equipment | |
| CN110414260B (en) | Data access method, device, system and storage medium | |
| CN115329871A (en) | Model training and model testing method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |