Google says adversaries are now “increasingly leveraging generative AI across multiple stages of the attack lifecycle,” from researching targets to drafting phishing messages and troubleshooting ...

On Thursday, Google announced that “commercially motivated” actors have attempted to clone knowledge from its Gemini AI ...

Threat actors have been spotted using complex techniques to figure out how mature large language models work, and using the ...

LLMs tend to lose prior skills when fine-tuned for new tasks. A new self-distillation approach aims to reduce regression and ...

Using The Water Dragon and Reunion as case studies, this paper applies Serafini’s multimodal text analysis framework to compare the Chinese and English covers from three perspectives: perception, ...

MIT researchers unveil a new fine-tuning method that lets enterprises consolidate their "model zoos" into a single, continuously learning agent.

Neel Somani has built a career that sits at the intersection of theory and practice. His work spans formal methods, mac ...

Abstract: We introduce Adversarial Sparse Teacher (AST), a robust defense method against distillation-based model stealing attacks. Our approach trains a teacher model using adversarial examples to ...

Reactive distillation is a process intensification technique that merges reaction and separation in one unit, reducing equipment count and energy use. It boosts conversion by continuously removing ...

Stephen Acabado receives funding from the Henry Luce Foundation and the National Science Foundation. But behind the spirit’s flash of marketing and growing popularity lies a rarely asked question: ...

Scientists demonstrate a process called "magic state distillation" in logical qubits for the first time, meaning we can now build quantum computers that are both error-free and more powerful than ...