Designing and Evaluating Language Models for Human Interaction

Explore the cutting-edge research on designing and evaluating language models for human interaction in this insightful talk by Mina Lee from Stanford University. Delve into innovative approaches for optimizing language models for real-world usage, including text infilling capabilities and the CoAuthor platform for capturing human-LM interaction in collaborative writing. Discover the HALIE framework for evaluating human-LM interaction beyond writing, and gain unique insights into language model capabilities regarding language, ideation, and collaboration. Learn about open challenges and future directions in this field from a rising star in AI research, recognized as one of MIT Technology Review's Korean Innovators under 35 in 2022. Benefit from Lee's expertise in building writing assistants and evaluating language models based on their ability to interact with humans and augment human capabilities.