Study on Automatic Coevolution between Cyber Attacks and Countermeasures Based on AI Technologies

2018 Fiscal Year Final Research Report

• Project/Area Number: 16K12439
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Information security
• Research Institution: Tokyo Denki University
• Principal Investigator: YAMAKI Hirofumi 東京電機大学, システムデザイン工学部, 教授 (10322166)
• Project Period (FY): 2016-04-01 – 2019-03-31
• Keywords: 人工知能 / 共進化学習 / 標的型攻撃 / マルウェア

Outline of Final Research Achievements

Based on the future image that targeted attacks are advanced using malware equipped with artificial intelligence technology, the attack contents are automatically synthesized and simulated in a form ahead of the attackers, and defense methods are automatically generated. We researched the technology to generate in.
Specifically, we constructed a network simulator to simulate malware attack and defense against it on a cloud computer, and constructed a system to reproduce various aspects by changing attack content and defense content by co-evolutionary calculation . In addition, we examined the attack model based on the planning calculation and the one that works the fraud through the dialogue with human being as a model of the artificial intelligence loaded malware.

Free Research Field

情報セキュリティ

Academic Significance and Societal Importance of the Research Achievements

本研究は，(1)マルウェアや攻撃対策システムを自律的なソフトウェアであるエージェントとしてモデル化し，(2)共進化計算を通じてマルウェアと攻撃対策システムとのあいだの「イタチごっこ」をシミュレーションすることにより(3)将来高度化すると予想されるマルウェアの攻撃に先んじて対策を構築するというものである．
この技術の開発により，新しい攻撃の出現を対策が後追いするというセキュリティ対策技術の宿命から脱し，攻撃者の先回りをして防御をするという，真の意味でプロアクティブなセキュリティ対策が実現する．