Algorithm model deployment and risk monitoring method and device, and computer device

Abstract

The invention relates to an algorithm model deployment and risk monitoring method and device, and computer equipment. The algorithm model deployment method comprises the steps: receiving an algorithmmodel file of a current algorithm model uploaded by a terminal; executing a related component library according to the algorithm model file and preset algorithms, and determining related components ofthe current algorithm model; and generating an algorithm package of the current algorithm model according to the algorithm model file and related components. The risk monitoring method comprises thesteps: sending an algorithm downloading request to an object management system, wherein the algorithm downloading request comprises identification information of a target risk analysis algorithm model, and deploying the risk analysis algorithm model stored in the object management system through the algorithm model deployment method; receiving an algorithm packet of the target risk analysis algorithm model returned by the object management system; and carrying out risk monitoring by loading and executing the algorithm packet. By adopting the algorithm model deployment and risk monitoring method, the service availability and the system robustness can be improved.

Description

technical field [0001] The present invention relates to the technical field of the Internet, in particular to a method, device, computer equipment and storage medium for algorithm model deployment and risk monitoring. Background technique [0002] With the application and development of artificial intelligence in the Internet field, a large number of algorithm models for different scenarios have been established, and how to flexibly deploy algorithm models has become an urgent research problem. [0003] In the currently adopted solution, the algorithm model file used for deployment is coupled with various execution components. In this way, the update of each small component will trigger the update of the entire algorithm system, which increases the version iteration frequency of the system. Each system update will cause the algorithm service to be stopped and restarted. During this period, the load capacity of the server will drop sharply, which will affect the specific busi...

AI Technical Summary

Problems solved by technology

[0003] In the currently adopted solution, the algorithm model file used for deployment is coupled with various execution components. In this way, the update of each small component will trigger the update of the entire algorithm system, which increases the version iteration frequency of the system. Each system update will cause the algorithm service to be stopped and restarted. During this period, the load capacity of the server will drop sharply, which will affect the specific business.

At the same time, due to the coupling of components, the modification of local code must be completed by modifying the entire system code, which may lead to system crash due to modification errors, which increases the risk of system maintenance and reduces the maintainability and robustness of the system.

Images