基於可信賴物聯網協調資產的電網韌性
該文提出了一個名為 EUREICA 的框架，透過利用本地電力市場協調可信賴的物聯網設備，提高電網面對網路攻擊的韌性。
Resilience of the Electric Grid against Cyberattacks through Trustable IoT-Coordinated Assets
The increasing reliance on IoT devices and distributed energy resources (DERs) in modern power grids, while beneficial for efficiency and sustainability, introduces new vulnerabilities to cyberattacks. This paper proposes a framework called EUREICA, which leverages local electricity markets (LEMs) and resilience scores to enhance grid resilience by enabling situational awareness and coordinated mitigation using trustable assets.
UnMarker：一種針對防禦性圖像水印的通用攻擊方法
本文指出，現有的防禦性圖像水印技術不足以抵禦深度偽造，並提出了一種名為 UnMarker 的通用攻擊方法，可以有效移除圖像中的水印，使得現有的深度偽造檢測技術失效。
방어적 이미지 워터마킹에 대한 보편적 공격: UnMarker
딥페이크 방지를 위해 고안된 방어적 이미지 워터마킹 기술이 'UnMarker'라는 새로운 공격 기술에 취약하며, 딥페이크 방지 대책을 재고해야 할 필요성을 제기합니다.
防御的画像電子透かしに対する普遍的な攻撃：UnMarker
ディープフェイク対策として期待されていた防御的画像電子透かしは、UnMarkerによってその有効性が疑問視されており、代替手段の検討が急務である。
UnMarker: A Universal Attack Method for Removing Defensive Watermarks from AI-Generated Images
Defensive watermarking, a technique used to identify AI-generated images, is not a reliable defense against deepfakes as it can be universally bypassed by a novel attack method called UnMarker.
Output Statistics of Random Binning Using Tsallis Divergence and Its Applications to Secure Communication
This research paper presents a novel framework for analyzing the output statistics of random binning using Tsallis divergence and demonstrates its application in achieving secure communication over wiretap channels.
SecONN: Protecting Optical Neural Networks from Thermal Fault Injection Attacks
SecONN is a novel framework that concurrently detects and protects against thermal fault injection attacks on optical neural networks, ensuring their reliability and security without compromising performance.
網站拒爬蟲入侵：探討內容創作者如何應對 AI 爬蟲的意識、能力與成效
雖然內容創作者迫切需要保護其作品免受 AI 爬蟲的侵害，但由於缺乏對現有技術方法（如 robots.txt）的認識、使用能力和實際操作的自主權，他們在很大程度上仍然處於弱勢，難以有效維護自身權益。
Somesite I Used To Crawl: AI 크롤러로부터 콘텐츠 제작자를 보호하기 위한 인식, 실행력 및 효능
본 논문에서는 콘텐츠 제작자들이 AI 크롤링으로부터 자신들의 콘텐츠를 보호하기 위해 robots.txt 및 능동적 차단과 같은 기술적 조치를 취하고자 하는 강한 열망이 있음에도 불구하고, 인식 부족, 기술적 장벽, 제한적인 효능으로 인해 어려움을 겪고 있음을 보여줍니다.
© 2024 by Linnk AI