2021 Fiscal Year Final Research Report
Development of a secure and non-fooled AI engine with multiple defense mechanisms
| Project/Area Number |
19K24357
|
|---|
| Research Category |
Grant-in-Aid for Research Activity Start-up
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
1001:Information science, computer engineering, and related fields
|
|---|
| Research Institution | Meijo University |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2019-08-30 – 2022-03-31
|
| Keywords | AIエンジン / セキュリティ |
|---|
| Outline of Final Research Achievements |
To realize a better society, the AI technology has been attracted attention. It is important to ensure the safety of AI to promote social implementation of AI. Therefore, this study conducted research and development to realize a secure AI system. This study revealed the threats of model extraction attacks and evasion attacks and researched the countermeasure methods. The proposed methods improved the security of AI system.
|
| Free Research Field |
情報セキュリティ
|
| Academic Significance and Societal Importance of the Research Achievements |
研究成果の学術的意義は，実デバイスに実装したAIエンジンでのいくつかの攻撃に対する脅威を明らかにした点と，開発した対策手法によって，AIエンジンに対する攻撃への安全性を向上できる点である。また，本研究で研究開発した手法によって， AIシステムのセキュリティを向上させることができるため，AIの社会実装を推進することができる点で社会的意義が大きい。
|
