A method and system for updating machine learning models based on self-learning

Abstract

The invention provides a self-learning-based machine learning model updating method. The method comprises the following steps of: creating an original machine learning model and deploying the model tobe online; detecting malicious attacks by the original machine learning model and storing the malicious attacks into a negative sample library; when the number of negative samples achieves a set threshold value, triggering a machine learning training task so as to create a new machine learning model; and carrying out model updating according to a set model updating strategy. The invention furthermore discloses a self-learning-based machine learning model updating system. The system comprises a negative sample library unit, a machine learning training unit and a model updating unit, wherein the negative sample library unit is used for storing detected malicious attacks; the machine learning training unit is used for triggering the machine learning training task when the number of the negative samples achieves the set threshold value so as to create the new machine learning model; and the model updating unit is used for carrying out model updating according to the set model updating strategy. According to the method and system, the early-stage sample collection pressure is decreased, and samples are collected through self-learning to be trained after the model is online; precision of the original model and precision of the new model are compared to decide whether the update the model without manual intervention; and the scenes which cannot carry out pushing and updating for external parts are overcome.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a method and system for updating a machine learning model based on self-learning. Background technique [0002] Machine learning (Machine Learning, ML) is a multi-field interdisciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. Specializes in the study of how computers simulate or implement human learning behaviors to acquire new knowledge or skills, and reorganize existing knowledge structures to continuously improve their performance. It is the core of artificial intelligence and the fundamental way to make computers intelligent. Its application pervades all fields of artificial intelligence. It mainly uses induction and synthesis rather than deduction. [0003] Machine learning is mainly divided into three different classes of learning methods: [0004] Supervised lea...

AI Technical Summary

Problems solved by technology

[0013] Aiming at the deficiencies in the above-mentioned prior art, the present invention provides a method and system for updating machine learning models based on self-learning to solve the problem of poor application of machine learning in network security in the prior art

Images