Protected: Can a Large Language Model Learn From Medical Clinical Notes Without Remembering Them?

Health systems are training language models on psychiatric notes to ease documentation burden. New research shows that even after de-identification, these models can retain enough to let an attacker tell whether a specific patient's notes were used to train them, and demonstrates a practical path to fix it.

This content is password-protected. To view it, please enter the password below.

Infomations

Time

Industry Spotlight

Chakra Pavan Kumar Kota

Senior Software Engineer at Jack in the Box

With over 11 years of combined academic and industry research experience, I currently serve as a Senior Software Engineer at Jack in the Box, a technology-driven organization in the United States. I hold a Master’s degree from the University of Missouri–Kansas City and am pursuing a Ph.D. in Artificial Intelligence and Data Science at the University of the Cumberlands. My professional expertise includes Artificial Intelligence, Machine Learning, Cybersecurity, and Cloud Computing.

Article & Blogs

Related Study