Real-time malicious sample detection method based on artificial intelligence processor and electronic device

Abstract

The invention provides a real-time malicious sample detection method based on an artificial intelligence processor and an electronic device, and the method comprises the steps: carrying out the resource partitioning of a global on-chip cache, a pulse array and a non-DNN calculation unit of the artificial intelligence processor according to a target network and a malicious sample detection mechanism; generating, by a compiler, an instruction file according to the resource division result; and when the malicious sample detection mechanism determines that the input data is a malicious sample, notifying the target neural network to stop calculation. According to the invention, the reasoning performance of an accelerator for executing a target network is not reduced, the risk that the system isattacked by malicious samples is avoided, the resource utilization rate of the artificial intelligence processor is greatly improved, the demand for memory bandwidth is reduced, and the detection algorithm is high in compatibility and good in adaptability.

Description

technical field [0001] The invention relates to the field of computing system and microprocessor security, in particular to a real-time malicious sample detection method based on an artificial intelligence processor and an electronic device. Background technique [0002] In recent years, the development speed of semiconductor chip manufacturing process has declined, which has slowed down the advancement of Moore's Law and gradually reached its physical limit. Today's computer systems rely on dedicated hardware accelerators to achieve better performance and energy efficiency. Computation of machine learning models, especially deep neural networks, is usually massively compute-intensive and memory-intensive, both of which require specialized hardware accelerators to improve the performance and energy efficiency of their execution. Academia and industry have made great efforts to this end. In academia, in 2014, researcher Chen Yunji of the Institute of Computing Technology, Ch...

AI Technical Summary

Problems solved by technology

[0006] In order to overcome the deficiencies of the existing artificial intelligence processor technology, the present invention provides a real-time malicious sample detection method and electronic device based on an artificial intelligence processor, which can simultaneously run a target neural network model and a malicious sample detection model (including a neural network malicious sample detection model). detection model, traditional machine learning malicious sample detection model, or a combination of the two), so that the target neural network model can perform reasoning efficiently, and also enable the malicious sample detection algorithm to effectively detect malicious sample attacks in real time

Images