Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks

no code implementations • 16 Oct 2023 • Erfan Shayegani, Md Abdullah Al Mamun, Yu Fu, Pedram Zaree, Yue Dong, Nael Abu-Ghazaleh

Large Language Models (LLMs) are swiftly advancing in architecture and capability, and as they integrate more deeply into complex systems, the urgency to scrutinize their security properties grows.

Adversarial Attack Federated Learning

DeepMem: ML Models as storage channels and their (mis-)applications

no code implementations • 17 Jul 2023 • Md Abdullah Al Mamun, Quazi Mishkatul Alam, Erfan Shaigani, Pedram Zaree, Ihsen Alouani, Nael Abu-Ghazaleh

In this paper, we propose a novel information theoretic perspective of the problem; we consider the ML model as a storage channel with a capacity that increases with overparameterization.

Data Augmentation