Modern Generative AI
Modern generative AI focuses on creating realistic and diverse content, such as text, images, and even molecules, using deep learning models like transformers and diffusion models. Current research emphasizes improving model safety and trustworthiness, including developing methods for "unlearning" undesirable information and creating tools to enhance user understanding of these complex systems. This rapidly advancing field has significant implications across various sectors, from accelerating pharmaceutical R&D and automating software development to enhancing creative content generation and improving educational tools.
Papers
September 17, 2024
August 8, 2024
March 21, 2024
February 28, 2024
December 19, 2023
March 21, 2023