AI for Women's Health with Dr. Umbereen Nehal, MD, MPH

Join us on June 4th to hear from Dr. Umbereen Nehal, MD, MPH on how AI can impact women's health! We will also have Michael Fahey returning for a second session on AI and de-identification of PHI.

Generative AI, like chatbots and LLMs, is the hot new innovation in healthcare. One study showed AI to have more "empathy" than human doctors online. Dr. Umbereen Nehal, founder of HER Heard and an MIT Sloan Fellow, will describe use of GenAI in women's health, especially when used as a health navigator and medical assistant. The session explores choices in building generative AI, especially the ways to reduce hallucinations or errors and how to build inclusive AI for personalized care to counter bias. With ethical and responsible use of AI we can accelerate towards a future of greater health equity and health justice.

About Dr. Nehal:

Dr. Nehal is a founder of a digital health venture incubated at MIT, inspired by her own patient journey and the stories of over 200 women. Her team is designing an AI-enabled FemTech solution to improve women’s health for common conditions like fibroids, menopause, endometriosis, and pelvic pain.

As an MIT Sloan Fellow with additional coursework via Stanford LEAD, Dr. Nehal has honed skills in AI-enabled big data analytics and human-centered design for healthcare.

A seasoned physician executive, a former Chief Medical Officer for 14 centers, a co-author of $1.8b in Medicaid value-based care reform, Dr. Nehal is skilled at public-private partnerships to advance cross-sector solutions for population health. She is a keynote speaker and was recognized by President Obama for her community organizing in marginalized communities. She is focused on systems approaches for improved health and well-being outcomes.

Our second presentation with Michael Fahey will be “More on Using LLMs and GenAI with Healthcare Data”

This presentation will be a follow up to the last Meetup presentation ‘Using LLMs and GenAI with Healthcare Data, HIPAA, and De-Identification’. You can see this past event here.

This presentation will be a follow up to the last Meetup presentation ‘Using LLM’s and GenAI with Healthcare Data, HIPAA, and De-Identification’. There is a link to that presentation here.
This session will explore options for Security and Privacy of PHI when working with Healthcare data and GenAI. In particular, the session will review requirements for hosting PHI, potential issues with using PHI with Private and/or hosted LLM’s, and setting up a HIPAA Environment. It will also review lower-cost options for LLM development and deployment for production including Quantization and other model reduction techniques. It will also review Fine-Tuning, Zero-Shot, One-Shot and Few-Shot Learning, Retrieval Augmented Generation, and more. It will explore analysis and deployment of Public LLM’s available on Hugging Face and other Model Hubs and using tools to measure model performance and data analysis.

Finally, the session will review the latest updates from NIST on the new Risk Management Framework for GenAI.

Although the session is focused on addressing the challenges of working with Healthcare Data and using low-cost options for LLM and GenAI development and deployment, it will also be useful for anyone outside of healthcare looking to understand Risk Management of GenAI or working on a budget looking for simpler options for learning to use LLM’s and GenAI.

Agenda:

5:30 - 6:00 Arrive, Food, Networking
6:00 - 7:00 Presentation and Discussion with Dr. Nehal
7:00 - 8:00 Presentation and Discussion with Michael Fahey
8:00 - 8:30 Networking and Open Discussion

There is a suggested optional donation of $5 to cover the costs of the event, which will be payable in cash or venmo at the event. Event costs are low, as the space is provided for free by Microsoft, but the Boston AI Group for Healthcare, Law, and Education does not have nor does it seek corporate sponsorship. This allows us to offer presentations free from any particular corporate interest. Microsoft does not moderate our events for content, nor are any of the views presented representative of Microsoft’s views.

RSVP is required, and a government issued ID is needed to enter the building. No one who has not RSVP'd will be able to attend, due to building security policies. Please ensure your meetup account includes your real first and last name to ensure no issues entering the building.