Advancing Responsible Large Language Models (LLMs) for Biomedicine and Healthcare

Are you interested in learning how large language models (LLMs) can address errors and biases in Electronic Health Record (EHR) data?

Join this NCI webinar to gain insights from Dr. Qi Long, who will explore how LLMs offer a promising solution to data issues, especially those stemming from incomplete information.

Dr. Long will share his team’s recent work in this space, including:

• mCodeGPT for extracting Minimal Common Oncology Data Elements (mCODE) from EHRs.
• SDoH-GPT for extracting Social Determinants of Health (SDoH) from unstructured data in EHRs.
• Multimodal Graph-LLM for predicting clinical events using both structured and unstructured EHR data. 

Additionally, he will discuss ongoing and planned research focused on developing rigorous statistical and machine learning methods to address various issues and biases with LLMs.