Foundation Models Tutorial and Why Not to Fine-Tune Them

Explore a comprehensive tutorial on Foundation Models presented by Ananya Kumar, a Stanford University PhD student. Dive into the development of better algorithms for pre-training and fine-tuning these models, with a focus on robustness and safety. Learn about the capabilities of Foundation Models and their adaptability for various tasks. Understand the potential risks and harms associated with these models, emphasizing responsible usage. Discover the interdisciplinary approach of Stanford's Center for Research on Foundation Models towards advancing and responsibly using these technologies. Gain insights into a specific project that demonstrates how fine-tuning can distort pre-trained features and underperform in out-of-distribution scenarios. Essential viewing for those interested in machine learning, AI models, and their real-world applications.

Foundation Models Tutorial, and Why Not to Fine Tune Them