Identifying and Understanding Cross-Class Features in Adversarial Training

Adversarial training AT has been considered one of the most effective methods for making deep neural networks robust against adversarial attacks, while the training mechanisms and dynamics of AT remain open research problems. In this paper, we present a novel perspective on studying AT through th...

OpenJDK: missing type safety checks for MethodHandle calls across class loaders, incorrect CVE-2013-5838 fix (Hotspot, 8151666)

An improper type safety check was discovered in the Hotspot component. An untrusted Java application or applet could use this flaw to bypass Java Sandbox restrictions...