Рет қаралды 538
2023 was a breakthrough year for all things Machine Learning, especially for generative use cases. The community saw models and frameworks released at a blistering pace thanks to innovative solutions to fit LLMs onto smaller and smaller devices (LoRA, PEFT, etc). The industry saw the rise of risk assessment frameworks and legislative actions, leading to new requirements for organizations to follow.
Our talk will cover various pairing of these state of the art (SOTA) models and techniques with the traditional security and identity tools like Bloodhound, Seatbelt, and Mythic. We’ll introduce the framework of “Offensive ML” as a discipline and work to use machine learning to support better security solutions. AI promises massive change across cybersecurity, but rather than speculate in the abstract, we’ll focus on practical solutions and demonstrations.