An abnormal group identification method, device and smart chip

Abstract

The present invention relates to an abnormal group identification method, device and intelligent chip. The abnormal group identification method includes: extracting characteristic information to be identified from monitored data intelligence; inputting the characteristic information to be identified into a target network model to obtain anomaly groups; extract the virtual identity information of multiple virtual identity users from the network identity data of different network platforms; based on the extracted virtual identity information and the real identity information included in the monitored data intelligence, establish the identification of each abnormal group in the abnormal group The association relationship between group members and target virtual identity users; the training method of the target network model includes: obtaining the initial network model; simplifying the convolution layer in the initial network model to obtain a simplified network model; using the preset training algorithm, training The data is input into the simplified network model for training to obtain the target network model. The solution of the invention can reduce the power consumption of the smart chip in the process of data processing and calculation.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to an abnormal group identification method, device and intelligent chip. Background technique [0002] With the continuous development of artificial intelligence technology, the demand for computing power of smart chips is increasing exponentially, and large computing power means that power consumption also increases. Moreover, with the development of neural network technology, the deep learning framework (Caffe) has been widely used. [0003] After training, the Caffe-based neural network model can process data such as images, voices, and texts to obtain the required recognition results, such as recognizing images to obtain image features, and recognizing voices to obtain control instructions. In the traditional neural network model, the amount of data that needs to be processed is increasing, so the energy efficiency of the smart chip in the data processing and calculation proc...

AI Technical Summary

Problems solved by technology

In the traditional neural network model, the amount of data that needs to be processed is increasing, so the energy efficiency of the smart chip in the data processing and calculation process is high, resulting in a large power consumption of the smart chip

Images