Abstracts: December 6, 2023

Join the newest episode of Abstracts, where leading researchers at Microsoft share their cutting-edge breakthroughs and achievements. In this episode, Xing Xie, a Senior Principal Research Manager of Microsoft Research Asia discusses “Evaluating General-Purpose AI with Psychometrics” with host Dr. Gretchen Huizinga. The discussion focuses on psychometrics, a subfield of psychology, and its role as a new evaluation method for general-purpose AI models. The paper demonstrates how psychometrics can help evaluate model performance and support consistent and reliable design of AI systems. This exciting field is driving the future of AI and promises to revolutionize how we approach evaluation and design.

To learn more about Evaluating General-Purpose AI with Psychometrics, check out the associated paper.

Listen to the full episode on Spotify.