About
Research & Development
SAFEMax: A New Approach to Efficiently Erase Data from Generative AI Models
TLDR: SAFEMax is a novel method for Machine Unlearning in diffusion models that efficiently makes them forget specific content by maximizing entropy in generated...
Research & Development
Unpacking Compositional Understanding in Vision-Language Models
TLDR: A study evaluated Diffusion Classifier, CLIP, and ViLT on their ability to understand how concepts combine (compositionality) in images. While Diffusion Classifier and...
Research & Development
MAPTA: Autonomous AI Agents Enhance Web Security Auditing
TLDR: MAPTA is a multi-agent AI system for autonomous web application penetration testing. Developed by researchers at University College London, it combines large language...
Research & Development
Simplifying Robot Intelligence with Primitive Actions
TLDR: The research introduces Primitive Embodied World Models (PEWM), a novel approach that breaks down complex robotic tasks into short, fundamental "primitive" motions. This...
Research & Development
Deconstructing Jailbreaks: A New Framework for Accurate LLM Security Assessment
TLDR: JADES is a new framework that accurately assesses LLM jailbreak attempts by breaking down harmful questions into weighted sub-questions, scoring sub-answers, and aggregating...
Research & Development
Train Once, Plan Anywhere: Advancing Robot Kinodynamic Motion Planning
TLDR: DiTree is a new framework that combines diffusion models with sampling-based planners to improve kinodynamic motion planning for robots. It uses learned diffusion...