Permissions based Automatic Android Malware Repair using Long Short Term Memory

롱 숏 텀 메모리를 활용한 권한 기반 안드로이드 말웨어 자동 복구

  • Wu, Zhiqiang (Dept. of Computer Science and Engineering, Hanyang University) ;
  • Chen, Xin (Dept. of Computer Science and Engineering, Hanyang University) ;
  • Lee, Scott Uk-Jin (Dept. of Computer Science and Engineering, Hanyang University)
  • 오지강 (한양대학교 컴퓨터공학과) ;
  • 천신 (한양대학교 컴퓨터공학과) ;
  • 이욱진 (한양대학교 컴퓨터공학과)
  • Published : 2019.01.16

Abstract

As malicious apps vary significantly across Android malware, it is challenging to prevent that the end-users download apps from unsecured app markets. In this paper, we propose an approach to classify the malicious methods based on permissions using Long Short Term Memory (LSTM) that is used to embed the semantics among Intent and permissions. Then the malicious method that is an unsecured method will be removed and re-uploaded to official market. This approach may induce that the end-users download apps from official market in order to reduce the risk of attacks.

Keywords