Distributed artificial intelligence calculation method, Internet of Things node and storage medium

Abstract

The invention discloses a distributed artificial intelligence calculation method, an Internet of Things node and a storage medium, and the method is used for a multi-hop network comprising a pluralityof Internet of Things nodes, and comprises the steps of obtaining at least one group of learning sample data related to a supervised artificial neural network model, sending the plurality of pieces of calculation task allocation information to the plurality of Internet of Things nodes respectively, completing calculation tasks at the Internet of Things nodes according to the learning sample data,the weight information and the activation function information to obtain calculation results, calculating error data according to the calculation results and the label data, and updating the weight information of each Internet of Things node according to the error data. By implementing the invention, the distributed artificial intelligence computing network composed of the Internet of Things nodes can learn features in the sample data, the recognition precision of the computing network is improved, and errors are reduced.

Description

technical field [0001] The invention relates to the technical field of distributed computing, in particular to a distributed artificial intelligence computing method, an Internet of Things node and a storage medium. Background technique [0002] With the development of IoT technology, there are more and more IoT nodes, and these IoT nodes are usually powered by batteries, so their computing power is often limited, and due to the limitation of battery life, these IoT nodes are not It is not suitable to independently undertake more complex artificial intelligence computing tasks based on supervised artificial neural network models, such as the computing tasks of artificial neural network-based artificial intelligence environmental monitoring models. Therefore, in order to avoid processing these computing tasks locally at the IoT nodes, the existing technicians often send the data collected by the IoT nodes to the cloud for processing, which will bring a lot of communication de...

AI Technical Summary

Problems solved by technology

[0002] With the development of IoT technology, there are more and more IoT nodes, and these IoT nodes are usually powered by batteries, so their computing power is often limited, and due to the limitation of battery life, these IoT nodes are not It is not suitable to independently undertake more complex artificial intelligence computing tasks based on supervised artificial neural network models, such as the computing tasks of artificial neural network-based artificial intelligence environmental monitoring models

Therefore, in order to avoid processing these computing tasks locally at the IoT nodes, existing technicians often send the data collected by the IoT nodes to the cloud for processing. Gateways or wireless communication modules, thus increasing hardware costs, while the computing power of IoT nodes is not fully utilized

Images