Computer security

Computer science

Malware

Artificial intelligence

Deep learning

Residual neural network

Journal of Computer Virology and Hacking Techniques

We have utilized two distinct models to identify the obscure or new sort of malware in this paper. GoogleNet and ResNet models are researched and tried which belong to two different platforms i.e. ResNet belongs to Microsoft and GoogleNet is the intellectual property of Google. Two sorts of datasets are utilized for training and validation the models. One of the dataset was downloaded from Microsoft which is the combination of 10,868 records and these records are binary records. These records are additionally isolated in nine diverse classes. Second dataset is considerate dataset and it contains 3000 benign files. The said datasets were initially in the form of EXE files and were changed over into opcode, after that changed over into images. We got a testing accuracy of 74.5% on GoogleNet and 88.36% precision on ResNet.

Analysis of ResNet and GoogleNet models for malware detection