Movatterモバイル変換


[0]ホーム

URL:


CN114764389B - Heterogeneous simulation test platform for joint learning systems - Google Patents

Heterogeneous simulation test platform for joint learning systems
Download PDF

Info

Publication number
CN114764389B
CN114764389BCN202110047867.2ACN202110047867ACN114764389BCN 114764389 BCN114764389 BCN 114764389BCN 202110047867 ACN202110047867 ACN 202110047867ACN 114764389 BCN114764389 BCN 114764389B
Authority
CN
China
Prior art keywords
data
information
client
load
model
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
Application number
CN202110047867.2A
Other languages
Chinese (zh)
Other versions
CN114764389A (en
Inventor
李振飞
王瑞扬
刘伟赫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xinao Xinzhi Technology Co ltd
Original Assignee
Xinao Xinzhi Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xinao Xinzhi Technology Co ltdfiledCriticalXinao Xinzhi Technology Co ltd
Priority to CN202110047867.2ApriorityCriticalpatent/CN114764389B/en
Publication of CN114764389ApublicationCriticalpatent/CN114764389A/en
Application grantedgrantedCritical
Publication of CN114764389BpublicationCriticalpatent/CN114764389B/en
Activelegal-statusCriticalCurrent
Anticipated expirationlegal-statusCritical

Links

Classifications

Landscapes

Abstract

Translated fromChinese

本发明公开了联合学习系统异构模拟测试平台,包括服务器、客户端和运行环境,所述服务器提供计算服务,作为网络的节点,存储并处理网络上的数据和信息,所述客户端与服务器相对应,为客户提供本地服务的程序,一般安装在普通的客户机上,需要与服务器互相配合运行,所述运行环境对虚拟环境模拟参数进行判断。该联合学习系统异构模拟测试平台,使用模拟、抽象的形式,在实际的联合学习应用场景中不同边缘设备的计算、通信、存储能力不同,根据已知的系统数据比如硬件计算能力、网络延迟、通信带宽等进行抽象、建模,构建一个和真实环境相似的虚拟系统异构测试平台,可以模拟不同的实际应用场景以验证不同的联合学习解决方案,实用性更强。

The present invention discloses a heterogeneous simulation test platform for a joint learning system, including a server, a client, and an operating environment. The server provides computing services, acts as a node of a network, stores and processes data and information on the network, and the client corresponds to the server, provides a program for local services to clients, is generally installed on an ordinary client, and needs to cooperate with the server to run, and the operating environment judges the virtual environment simulation parameters. The heterogeneous simulation test platform for a joint learning system uses simulation and abstract forms. In the actual joint learning application scenario, different edge devices have different computing, communication, and storage capabilities. Based on known system data such as hardware computing power, network latency, and communication bandwidth, abstraction and modeling are performed to build a virtual system heterogeneous test platform similar to the real environment. Different actual application scenarios can be simulated to verify different joint learning solutions, which is more practical.

Description

Heterogeneous simulation test platform of joint learning system
Technical Field
The invention relates to the technical field of a joint learning system, in particular to a heterogeneous simulation test platform of the joint learning system.
Background
Joint learning refers to a series of algorithms, but can collect parameters of a model, a server coordinates edge devices to participate in learning, each edge device has learning data, each edge device learns a local model by using own data, and the own parameters are encrypted or not encrypted and uploaded to the server, the server averages or weighted averages the collected parameters and broadcasts the parameters to each edge device, the joint learning is a distributed machine learning method, can learn a large amount of scattered data stored on a mobile phone and other devices, is a more generalized implementation of introducing codes into data instead of introducing data into codes, solves basic problems about privacy, ownership, data position and the like, can realize a more intelligent model, has lower delay and lower power consumption, and also ensures privacy.
Joint learning may allow mobile devices to co-learn a shared predictive model while maintaining all learning data on the device, thereby decoupling machine learning and cloud storage data, downloading the current model in the device, refining it by learning data on the device, and then assembling changes into a centralized small update. Only this update will be sent to the cloud and transmitted using encryption techniques. It will immediately average with other users' updates at the cloud to improve the sharing model. All the learning data are still kept on the local equipment and are not sent to the cloud for storage, the heterogeneous systems are two or more systems with different frameworks, the two systems (such as the sql database, the postgres database and the like) cannot be directly communicated, the data can be mutually accessed through interfaces and the like, and the systems consisting of a plurality of different framework systems are heterogeneous systems, and can not simulate different practical application scenes to verify the defects of different joint learning solutions due to different computing, communication and energy storage capacities of different edge devices in the practical joint learning application scenes.
Disclosure of Invention
The invention aims to provide a heterogeneous simulation test platform of a joint learning system, which aims to solve the problem that a virtual system heterogeneous test platform similar to a real environment cannot be constructed in the background technology.
In order to achieve the aim, the invention provides the technical scheme that the heterogeneous simulation test platform of the joint learning system comprises a server, a client and an operating environment,
The server provides computing service as a node of the network, and stores and processes data and information on the network;
The client corresponds to the server, and a program for providing local service for the client is generally installed on a common client and needs to be matched with the server for running;
and the running environment judges the simulation parameters of the virtual environment.
Preferably, the server may declare various parameters of the client, which send data and information to the client via an initialized client list/parameters, load data (detection), load model, and learn.
Preferably, the client sequentially performs environment initialization, load data and load model, the environment initialization data and information are sent to the operation environment, and the initial values of each round of operation environment are the same and normally distributed.
Preferably, the virtual simulation environment parameters in the operating environment include operating speed, net speed, communications and hardware.
Preferably, the loading data (detection) distributes the data set to the loading data of the client, and the data and information of the loading data are sent to the loading model.
Preferably, the data and information of the load model are split into two lines, one for local learning and one for local testing.
Preferably, the data and information of the loading model are sent to learning to judge, when judging as 'Y', the model is distributed to local learning, and when judging as 'N', the data and information are sent to global test.
Preferably, the locally learned data and information is sent to the aggregate after loss/accuracy/model parameters, and the aggregate data and information is sent to the global test.
Preferably, the global test sends data and information to the local test after passing through the global model.
Preferably, the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And secondly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step three, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Compared with the prior art, the invention has the beneficial effects that:
The heterogeneous simulation test platform of the joint learning system performs abstraction and modeling according to a known system, uses simulation and abstraction modes to simulate and abstract the problem of limitation of hardware, network, computing resources and the like in a joint learning real environment, solves the heterogeneous problem of the joint learning system, performs abstraction and modeling according to known system data such as hardware computing capacity, network delay, communication bandwidth and the like in an actual joint learning application scene, constructs a virtual system heterogeneous test platform similar to the real environment, can simulate different actual application scenes to verify different joint learning solutions, has stronger practicability, and realizes combination and sharing of data information resources, hardware equipment resources and manpower resources among different databases. The key point is that a global data mode or a global external view is established based on a local database mode, meanwhile, collected data also supports access to historical data, a user performs decision-supported query through a unified data interface provided by a data warehouse, the global mode is especially important for establishing an advanced decision support system, a joint learning system can collect parameters of a model, a server coordinates edge devices to participate in learning, each edge device has learning data, each edge device learns a local model by utilizing own data, the own parameters are encrypted or not encrypted and uploaded to the server, the server performs average or weighted average on the collected parameters and broadcasts the parameters to each edge device, the joint learning can generate a more intelligent model, lower delay and lower power consumption, and simultaneously, the privacy of the user is ensured, the cloud forms collaborative update for a shared model, unhook the demand of machine learning and cloud storage data, make the model more smart, lower in delay and more energy-saving, protect the privacy of users from being threatened, besides realizing the update of the shared model, the users can use the improved model immediately, the obtained experience can be different according to different personal use modes, relevant data is accessed through the internet of things, then model learning, model updating and calculation storage are carried out locally, a series of aggregation calculation and processing are carried out on the models provided by each user, and the combined global model is issued to each user, iterated back and forth until a better model is learned, thereby being convenient for the users to call better and share value, being applicable to edge devices with different calculation, communication and storage capacities, having stronger functionality and practicability, the operation is convenient, and the combined learning effect is better.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention provides a technical scheme that the heterogeneous simulation test platform of a joint learning system comprises a server, a client and an operating environment,
The server provides computing service as a node of the network, and stores and processes data and information on the network;
The client corresponds to the server, and a program for providing local service for the client is generally installed on a common client and needs to be matched with the server for running;
The running environment judges the simulation parameters of the virtual environment.
Further, the server may declare various parameters of the client, load data (detection), load model, and learn, declare various parameters of the client to send data and information to the client via an initialized client list/parameters.
Further, the client sequentially performs environment initialization, load data and load models, the data and information of the environment initialization are sent to the operation environment, and the initial values of each round of operation environment are the same and normally distributed.
Further, virtual simulation environment parameters in the operating environment include operating speed, net speed, communications, and hardware.
Further, the loading data (detection) distributes the data set to the loading data of the client, and the data and information of the loading data are sent to the loading model.
Further, the data and information of the load model are split into two lines, one to local learning and one to local testing.
Further, data and information of the loaded model are sent to learning to be judged, when the judgment is "Y", the model is distributed to local learning, and when the judgment is "N", the data and the information are sent to global test.
Further, the locally learned data and information is sent to the aggregate after loss/accuracy/model parameters, and the aggregate data and information is sent to the global test.
Further, the global test sends data and information to the local test after passing through the global model.
Embodiment one:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And secondly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step three, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Embodiment two:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And step two, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And thirdly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Embodiment III:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And secondly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step three, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, the locally learned data and information are transmitted to the total after loss/accuracy/model parameters, and the total data and information are transmitted to the global test.
And seventhly, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Embodiment four:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And step two, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And thirdly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, the locally learned data and information are transmitted to the total after loss/accuracy/model parameters, and the total data and information are transmitted to the global test.
And seventhly, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Fifth embodiment:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And secondly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And thirdly, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And step four, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Example six:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And secondly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And thirdly, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fourthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And fifthly, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Embodiment seven:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
And secondly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And thirdly, dividing the data and information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Example eight:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
the server can declare various parameters of the client, load data (detection), load model and study, and the various parameters of the client can send data and information to the client through the initialized client list/parameters.
The virtual simulation environment parameters in the operation environment comprise operation speed, net speed, communication and hardware
And thirdly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed. .
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Example nine:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
The virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And secondly, the server can declare various parameters of the client, load data (detection), load a model and learn, and the various parameters of the client can be declared to send data and information to the client through an initialized client list/parameter.
And thirdly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And seventhly, sending the locally learned data and information to the total after loss/accuracy/model parameters, and sending the total data and information to the global test.
And step eight, the global test sends the data and the information to the local test after passing through the global model.
Example ten:
the heterogeneous simulation test platform of the joint learning system comprises the following steps:
The virtual simulation environment parameters in the running environment comprise running speed, net speed, communication and hardware.
And secondly, the server can declare various parameters of the client, load data (detection), load a model and learn, and the various parameters of the client can be declared to send data and information to the client through an initialized client list/parameter.
And thirdly, sequentially carrying out environment initialization, load data and load models by the client, and sending the data and information of the environment initialization to the operation environment, wherein the initial values of each round of operation environment are the same, and the data and the information are normally distributed.
And step four, loading data (detecting) and distributing the data set to load data of the client, and sending the data and information of the load data to a load model.
And fifthly, dividing the data and the information of the load model into two lines, wherein one line is sent to local learning and the other line is sent to local testing.
And step six, data and information of the loaded model are sent to learning to be judged, when the judgment is Y, the model is distributed to local learning, and when the judgment is N, the data and the information are sent to global test.
And step seven, the global test sends the data and the information to the local test after passing through the global model.
And step eight, the locally learned data and information are sent to the aggregate after loss/accuracy/model parameters, and the aggregate data and information are sent to the global test.
Finally, it should be noted that the above description is only for illustrating the technical solution of the present invention, and not for limiting the scope of the present invention, and that the simple modification and equivalent substitution of the technical solution of the present invention can be made by those skilled in the art without departing from the spirit and scope of the technical solution of the present invention.

Claims (4)

Translated fromChinese
1.联合学习系统异构模拟测试平台,包括服务器、客户端和运行环境,其特征在于:1. The heterogeneous simulation test platform of the joint learning system includes a server, a client and an operating environment, and is characterized by:所述服务器提供计算服务,作为网络的节点,存储并处理网络上的数据和信息;The server provides computing services, acts as a node on the network, and stores and processes data and information on the network;所述客户端与服务器相对应,为客户提供本地服务的程序,一般安装在普通的客户机上,需要与服务器互相配合运行;The client corresponds to the server and is a program that provides local services to customers. It is usually installed on a common client and needs to work in conjunction with the server.所述运行环境对虚拟环境模拟参数进行判断;The operating environment determines the virtual environment simulation parameters;所述服务器可以声明客户端的各种参数、加载数据、加载模型和学习,所述声明客户端的各种参数经初始化客户端列表/参数将数据和信息发送至客户端;The server can declare various parameters of the client, load data, load models and learn, and the server can declare various parameters of the client and send data and information to the client by initializing the client list/parameters;所述客户端依次进行环境初始化、负载数据和负载模型,所述环境初始化的数据和信息发送至运行环境,所述运行环境每一轮的初始值一样,正态分布;The client performs environment initialization, load data and load model in sequence, and the environment initialization data and information are sent to the operating environment. The initial value of each round of the operating environment is the same and normally distributed;所述负载模型的数据和信息分成两条线路,一条送往本地学习,一条送往本地测试;所述本地学习的数据和信息经过损失/准确度/型号参数后发送至总计,所述总计的数据和信息发送至全局测试;所述全局测试经全局模型后将数据和信息发送至本地测试;The data and information of the load model are divided into two lines, one is sent to local learning, and the other is sent to local testing; the data and information of the local learning are sent to the total after passing through the loss/accuracy/model parameters, and the data and information of the total are sent to the global test; the global test sends the data and information to the local test after passing through the global model;所述运行环境中的虚拟模拟环境参数包括:运行速度、净速度、通信和硬件。The virtual simulation environment parameters in the operating environment include: operating speed, net speed, communication and hardware.2.根据权利要求1所述的联合学习系统异构模拟测试平台,其特征在于:所述加载数据分发数据集到客户端的负载数据,所述负载数据的数据和信息发送至负载模型。2. The heterogeneous simulation test platform of the joint learning system according to claim 1 is characterized in that: the loading data distributes the data set to the load data of the client, and the data and information of the load data are sent to the load model.3.根据权利要求1所述的联合学习系统异构模拟测试平台,其特征在于:所述加载模型的数据和信息发送至学习进行判断,当判断为“Y”时,模型分发至本地学习,当判断为“N”时,数据和信息发送至全局测试。3. According to the heterogeneous simulation test platform of the joint learning system in claim 1, it is characterized in that the data and information of the loaded model are sent to the learning for judgment. When the judgment is "Y", the model is distributed to local learning. When the judgment is "N", the data and information are sent to the global test.4.根据权利要求1-3任一项所述的联合学习系统异构模拟测试平台,其特征在于:其步骤如下:4. The heterogeneous simulation test platform for the joint learning system according to any one of claims 1 to 3, characterized in that the steps are as follows:步骤一:服务器可以声明客户端的各种参数、加载数据、加载模型和学习,声明客户端的各种参数经初始化客户端列表/参数将数据和信息发送至客户端;Step 1: The server can declare various parameters of the client, load data, load models and learn, declare various parameters of the client, and send data and information to the client after initializing the client list/parameters;步骤二:客户端依次进行环境初始化、负载数据和负载模型,环境初始化的数据和信息发送至运行环境,运行环境每一轮的初始值一样,正态分布;Step 2: The client performs environment initialization, load data, and load model in sequence. The data and information of environment initialization are sent to the operating environment. The initial values of each round of the operating environment are the same and normally distributed.步骤三:运行环境中的虚拟模拟环境参数包括:运行速度、净速度、通信和硬件;Step 3: The virtual simulation environment parameters in the running environment include: running speed, net speed, communication and hardware;步骤四:加载数据分发数据集到客户端的负载数据,负载数据的数据和信息发送至负载模型;Step 4: Load the data distribution data set to the client's load data, and send the data and information of the load data to the load model;步骤五:负载模型的数据和信息分成两条线路,一条送往本地学习,一条送往本地测试;Step 5: The data and information of the load model are divided into two lines, one for local learning and the other for local testing;步骤六:加载模型的数据和信息发送至学习进行判断,当判断为“Y”时,模型分发至本地学习,当判断为“N”时,数据和信息发送至全局测试;Step 6: The data and information of the loaded model are sent to the learning for judgment. When the judgment is "Y", the model is distributed to the local learning. When the judgment is "N", the data and information are sent to the global test.步骤七:本地学习的数据和信息经过损失/准确度/型号参数后发送至总计,总计的数据和信息发送至全局测试;Step 7: Locally learned data and information are sent to the aggregate after loss/accuracy/model parameters, and the aggregated data and information are sent to the global test;步骤八:全局测试经全局模型后将数据和信息发送至本地测试。Step 8: Global test sends data and information to local test after passing through the global model.
CN202110047867.2A2021-01-142021-01-14 Heterogeneous simulation test platform for joint learning systemsActiveCN114764389B (en)

Priority Applications (1)

Application NumberPriority DateFiling DateTitle
CN202110047867.2ACN114764389B (en)2021-01-142021-01-14 Heterogeneous simulation test platform for joint learning systems

Applications Claiming Priority (1)

Application NumberPriority DateFiling DateTitle
CN202110047867.2ACN114764389B (en)2021-01-142021-01-14 Heterogeneous simulation test platform for joint learning systems

Publications (2)

Publication NumberPublication Date
CN114764389A CN114764389A (en)2022-07-19
CN114764389Btrue CN114764389B (en)2025-05-16

Family

ID=82363675

Family Applications (1)

Application NumberTitlePriority DateFiling Date
CN202110047867.2AActiveCN114764389B (en)2021-01-142021-01-14 Heterogeneous simulation test platform for joint learning systems

Country Status (1)

CountryLink
CN (1)CN114764389B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN115994367A (en)*2022-12-092023-04-21北京知道创宇信息技术股份有限公司Threat information system and threat information model training method

Citations (2)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN108183826A (en)*2017-12-292018-06-19江南大学A kind of Multiscale Fusion network simulation duty mapping method under isomerous environment
CN109947954A (en)*2018-07-092019-06-28北京邮电大学 Multi-task collaborative identification method and system

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
US11488054B2 (en)*2017-12-062022-11-01Google LlcSystems and methods for distributed on-device learning with data-correlated availability
CN111680798A (en)*2019-03-112020-09-18人工智能医生股份有限公司Joint learning model system and method, apparatus, and computer-readable storage medium
CN110942154B (en)*2019-11-222021-07-06深圳前海微众银行股份有限公司 Data processing method, device, device and storage medium based on federated learning
CN111625361B (en)*2020-05-262022-11-01华东师范大学 A joint learning framework based on cloud server and IoT device collaboration

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN108183826A (en)*2017-12-292018-06-19江南大学A kind of Multiscale Fusion network simulation duty mapping method under isomerous environment
CN109947954A (en)*2018-07-092019-06-28北京邮电大学 Multi-task collaborative identification method and system

Also Published As

Publication numberPublication date
CN114764389A (en)2022-07-19

Similar Documents

PublicationPublication DateTitle
Qu et al.ChainFL: A simulation platform for joint federated learning and blockchain in edge/cloud computing environments
CN113505882B (en)Data processing method based on federal neural network model, related equipment and medium
CN114978908B (en)Evaluation and operation method and device for computing power network node
CN105959347B (en)A kind of digital collaborative simulation system communication means based on P2P network
CN113435534A (en)Data heterogeneous processing method and device based on similarity measurement, computer equipment and computer readable storage medium
CN104754024A (en) A mobile Internet-based software and hardware integration method and architecture
CN112199154A (en)Distributed collaborative sampling central optimization-based reinforcement learning training system and method
CN118606081B (en) Method, device, equipment and storage medium for determining computing nodes
CN104866650B (en) A method and apparatus for creating an abstract device
Li et al.CEBOW: A Cloud‐Edge‐Browser Online Web3D approach for visualizing large BIM scenes
CN114764389B (en) Heterogeneous simulation test platform for joint learning systems
Zhang et al.Cellular traffic offloading via link prediction in opportunistic networks
CN113487041B (en)Transverse federal learning method, device and storage medium
CN114500428A (en)Navigation sharing method and device, electronic equipment, storage medium and program product
CN111368060B (en)Self-learning method, device and system for conversation robot, electronic equipment and medium
CN114153880A (en) Data cache control method, electronic device and storage medium
CN111224824B (en) An edge autonomous model construction method
Jia et al.Construction of SIoT CG virtual reality framework under ubiquitous clouds environment
CN109271371A (en)A kind of Distributed-tier big data analysis processing model based on Spark
CN110061863B (en) A Fairness-Based Distributed Multitasking Crowd Sensing Method in Sparse Networks
CN117033997A (en)Data segmentation method, device, electronic equipment and medium
CN113191733A (en)Project multi-party collaborative management system and method based on Internet
Liang et al.Online multi-skilled task assignment on road networks
Hou et al.A distributed deployment algorithm of process fragments with uncertain traffic matrix
Rajkumar et al.Moving to the cloud, fog, and edge computing paradigms: Convergences and future research direction

Legal Events

DateCodeTitleDescription
PB01Publication
PB01Publication
SE01Entry into force of request for substantive examination
SE01Entry into force of request for substantive examination
TA01Transfer of patent application right

Effective date of registration:20250401

Address after:065001 China (Hebei) Pilot Free Trade Zone Daxing Airport Area Langfang Airport Economic Zone Hangyidao Free Trade Zone Science and Technology Innovation Base 2101, Langfang City, Hebei Province

Applicant after:Xinao Xinzhi Technology Co.,Ltd.

Country or region after:China

Address before:100020 10th floor, Motorola building, 1 Wangjing East Road, Chaoyang District, Beijing

Applicant before:ENNEW DIGITAL TECHNOLOGY Co.,Ltd.

Country or region before:China

TA01Transfer of patent application right
GR01Patent grant
GR01Patent grant

[8]ページ先頭

©2009-2025 Movatter.jp