Controlling the Hallucinating Doctor by Dr. Demi Plessas

From the venue:

Thursday, March 27 - 6:00 PM - Zoom

Artificial intelligence, particularly large language models (LLMs), offers immense potential for improving efficiency and decision-making in various industries, including healthcare and the public sector. However, the risk of AI systems producing biased, inaccurate, or "hallucinated" information raises concerns about their use in critical situations. In this talk, we will present our framework for developing and deploying trustworthy AI systems in healthcare operations, administration, and population health, ensuring safety and effectiveness.

Dr. Demi Plessas is a distinguished data scientist, mathematician, and engineering manager with over a decade of experience conceptualizing, researching, building, deploying, maintaining, and managing ML and AI models, systems, and platforms across industries. Currently the Principal Data Scientist at Edifecs, they serve as the AI/ML architect for the company, leading innovative projects in healthcare AI with a focus on trustworthy, equitable and transparent models for critical applications like risk adjustment, provider performance scoring, and clinical documentation. Dr. Plessas has advised top government agencies and health plans on trustworthy AI, including CMS and the VA. A talented artist as well, their AI art has won multiple awards, placing fi rst and top 10 in competitive competitions. In recognition of their expository writing, they were awarded the prestigious Carl B. Allendoerfer Award from the Mathematical Association of America in 2023. Dr. Plessas is dedicated to advancing the fi eld of trustworthy AI to benefit society.