Generative AI & LLMs in Health & Medicine | Emerging Health Applications and Opportunities

Рет қаралды 2,498

7 ай бұрын

In this livestream, the National Academy of Medicine (NAM) Leadership Consortium: Collaboration for a Learning Health System (LC) explored the issues, activities, opportunities, and guideposts related to rapidly developing large language models (LLMs)/generative artificial intelligence (AI) tools for use in health and medicine.
The conversation brought together key players, such as health professionals, technology developers, and government stakeholders, to share perspectives on the key needs that can catalyze the basis of a shared agenda.
The meeting addressed developments in generative AI, specifically LLMs, and focused on policy considerations for health care, medicine, and public health. Ultimately, the meeting sought to spur collaborative stakeholder communication and create a framework for rapid engagement on emerging issues.
Learn more about the LC’s work on AI and health here: nam.edu/programs/value-scienc...

Пікірлер: 4

@victoriaoyekanmi872 6 ай бұрын

Very amazing.

@guybisschops5995 6 ай бұрын

GenAI for medicine has the potential to revolutionize the approach to diseases like Alzheimer's. While there are criticisms, the high cost and limited effectiveness of traditional drugs make it essential to explore alternative approaches. Personalizing treatment based on an individual's genetics and specific needs can lead to more targeted and potentially more effective interventions. It's important to consider the long-term benefits and cost-effectiveness of such approaches in the context of improving the quality of life for individuals with chronic diseases like Alzheimer's.

@guybisschops5995 7 ай бұрын

Must we be finetuned or not...? It depends on what are you will doing with the model...Alzheimer's disease is based on a hypothesis...not science...in a finetuned model of Alzheimer's disease with EHRs...you shall see other outcomes than in a big LLM with T parameters without finetuning on the amyloid hypothesis...we see a lot of hallucinations (unpredictable errors or inaccuracies)...your model needs clinical context ( clinical reasoning) it is different with other reasoning...in breast cancer AI shall fail on TNDC without Neural-Symbolic LM or symbolic reasoning...