Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore a comprehensive lecture on watermarking techniques for Large Language Models (LLMs) presented by Scott Aaronson from UT Austin and OpenAI. Delve into the innovative statistical watermarking scheme developed at OpenAI for LLM outputs. Gain insights into the broader context of theoretical and empirical research on LLM watermarking conducted over the past year. Examine various approaches to the AI attribution problem and understand the challenges associated with deploying watermarking solutions. Investigate the unresolved technical issues surrounding the development of text watermarking methods resistant to translation, paraphrasing, and similar attacks. Enhance your understanding of this crucial aspect of AI security and attribution in this informative 75-minute talk from the Simons Institute.
