Building On Foundation Models? Ensure They Are Trustworthy.

Almost a year ago to the day, the Center for Research on Foundation Models (CRFM) – which was then a new initiative of the Stanford Institute for Human-Centered Artificial Intelligence (HAI) – held a virtual workshop on Foundation Models. They chose "Foundation" as the name for these models as they entail training one model on a huge amount of data, then adapting it to many applications. The data can be text, images, speech, and more. The tasks that such models can perform include – but are not limited to – answering questions, analyzing sentiments, extracting information form text, labeling images, and recognizing objects. These Foundation Models use self-supervised learning, and they work because they can effectively apply knowledge learned in one task to another task.