Harmful website detection method based on generative adversarial network and deep learning

Abstract

The invention relates to a harmful website detection method and device based on a generative adversarial network and deep learning, and a readable medium. The method comprises the following steps: obtaining a plurality of first website snapshots comprising harmful websites through a crawler; inputting the first website snapshots as data of a generative adversarial network to obtain a plurality of simulated website snapshots; combining the simulated website snapshots with second website snapshots of a plurality of normal websites to obtain a training set; and finely tuning the convolutional neural network for training, and training the training set through the convolutional neural network to obtain a detection model for the harmful website. The invention has the beneficial effects that the required labeled data volume is greatly reduced, the time and computing resources required for training are also greatly reduced, and the model proposed in the technology has better adaptability for new personalized small sample tasks, can generate a reliable deep learning model in a short time, and improves the practical value of the system.

Description

technical field [0001] The invention relates to the field of computer security, in particular to a method, device and medium for detecting harmful websites based on generative confrontation networks and deep learning. Background technique [0002] With the continuous development of my country's Internet technology, the mining of harmful information on the Internet, which is widely used at present, usually needs to be analyzed and judged in combination with the massive control plane of the communication network and the online log data of the user plane. As the domestic privacy protection requirements are getting higher and higher, Especially under the requirements of operators to protect customer privacy, based on big data and AI, machine learning and deep learning technology, without involving user privacy, by generating massive Internet website snapshot data, the analysis and detection based on deep learning is more It has increasingly become an urgent means for the detection...

AI Technical Summary

Problems solved by technology

[0003] Existing models for identifying harmful information on the Internet often require a large amount of labeled data for training, but in practical applications, it is difficult to obtain a large amount of labeled training information for a specific field or user group, which leads to the use of Models trained with small sample data often fail to achieve the desired results. In addition, training a mature deep machine learning model often consumes a lot of time and computing resources, resulting in a significant increase in training costs for new tasks.

Images