Focuses on the development and evaluation of large pretrained models or foundational models (such as large language models) that process and integrate multiple forms of data (e.g., text, audio, video) to perform tasks that require a holistic understanding of diverse inputs.