Distributed AI system

Abstract

A distributed AI system can be a DDAI system and is used for flexibly and efficiently solving artificial intelligence application tasks. The DDAI system comprises: a registration unit used for registration when a component is dynamically accessed to the DDAI system or logout when the component is disconnected from the DDAI system; the task plan management unit that is used for planning and managing distributed AI tasks according to the characteristics of the accessed components; the task interaction unit that is used for interacting information between accessed components; the task execution unit that is used for executing the distributed AI subtasks by the accessed components so as to finish the distributed AI tasks; the standardization unit that is used for enabling the DDAI system to correspond to a unified space, and the unified space comprises a unified feature space and a unified marking space, wherein the component can be independent physical equipment or a cloud virtual node; and each assembly may carry one or more of the units.

Description

technical field [0001] The present application relates to the field of artificial intelligence (AI), in particular to a distributed artificial intelligence AI system. Background technique [0002] Machine learning systems are the most important branch of AI systems. Distributed machine learning (distributed machine learning, DML) system is a commonly used system for processing large-scale artificial intelligence application tasks. Traditional distributed machine learning systems are centralized systems that use computing clusters to obtain predictive models by training massive user data. Such a centralized system often requires intensive computing resources, and a large amount of user data is uploaded to the cloud for storage, which can easily cause privacy and security issues. [0003] In order to solve the above privacy protection problems, the federated learning (FL) system recently proposed an interactive mode between the local and the cloud, which protects the privacy...

AI Technical Summary

Problems solved by technology

However, for terminal devices, it is often faced with the fact that devices are connected to the network at any time and disconnected from the network, and the features collected by heterogeneous devices belong to different feature spaces. Even each terminal device faces different AI tasks and belongs to different label spaces. question

The above two AI systems are difficult to meet the needs of completing artificial intelligence application tasks under the conditions of dynamic connection and disconnection of the above-mentioned devices and heterogeneous devices

Furthermore, both of the above two AI systems rely on a central node for global synchronization, which will cause a large communication overhead, greatly affecting the flexibility and efficiency of solving AI tasks

Images