Federal machine learning migration method and system implemented for intelligent mobile terminal

Abstract

The system comprises a plurality of devices (1) distributed at different addresses, and an intelligent mobile terminal (2). Each device (1) comprises a data cleaning module (3), a data sensing module (6) and a data reading module (11); a data training fusion sub-module (4) is provided, wherein the data training fusion sub-module (4) is arranged on a part of the devices (1); a data training combination module (5) is provided, wherein the data training combination module (5) is arranged on the intelligent mobile terminal (2); all the devices (1) comprise single-machine storage modules (7), local data storage modules (8) are arranged on the devices (1) provided with the data training fusion sub-modules (4), and global data storage modules (9) are arranged on the devices (1) provided with the data training combination modules (5). According to the invention, a standardized interface mode is adopted for the data communication connection of the equipment and the mobile equipment, so that the operation module can be simultaneously connected with a plurality of systems needing federated machine learning, the data training cost is lower, and the portability of the system is improved.

Description

technical field [0001] The invention relates to the technical field of computer algorithms, in particular to a federated machine learning migration method and system for intelligent mobile terminals. Background technique [0002] At present, with the advancement of science and technology, scientific and technological manufacturing and automatic control have entered the era of intelligence from the previous braking. In the era of intelligence, in order to achieve artificial intelligence control, it is usually necessary to obtain sufficient knowledge in advance, that is, the corresponding parameter input and corresponding parameter output to form a mapping relationship, and based on the parameter input and corresponding parameter output, through the data Training methods and learning methods, so as to obtain a wider input and output mapping relationship, and then provide it to the control device in order to realize intelligent control. In order to obtain this mapping relations...

AI Technical Summary

Problems solved by technology

[0010] 1. In the existing technology, the reinforcement learning training model usually uses the data collected by itself for learning, optimization and control, and the knowledge of the reinforcement learning training model is also transferable. However, these systems for knowledge transfer learning are usually specified One device is the control device, and other working devices are used as training module providers. However, if the uniquely determined controller fails, the knowledge transfer of the reinforcement learning training model will fail.

[0011] 2. In the existing technology, centralized computing is mainly used for data learning and training, that is, when performing federated machine learning and training on the data of multiple devices, one of the computing devices is used for computing. When the data size is large, or the data dimension is relatively large, it takes a long time to complete the data training, or the processor requirements are very high, which will lead to an increase in processing costs, and when the computing processing device is set on a certain device, it will also increase the processing cost. The various requirements for equipment are strengthened, which is not conducive to the improvement and integration of the system

[0012] 3. In the prior art, when carrying out various trainings of data recording, the docking interfaces of the data acquisition module and the data training operation module of the equipment are various, and the data format is not unified and standardized, that is, there is no standardized interface, so the system can be transplanted poor sex

[0013] 4. Without considering the size and quantity of the data, when all the data is directly trained to obtain the model, it is easy to cause the amount of data to be too large, so that on the one hand, the amount of data calculation is large and the data calculation is difficult; at the same time, the large amount of data is easy Lead to inaccurate data training model

But so far, there is no effective way to solve the above technical problems in the prior art

Images