Healthcare Text Analytics in the Era of Large Language Models

HealTac2024 Tutorial

WEDNESDAY, JUNE 12TH 2024 14:00-17:00

InfoLab21, Floor B, Room B79

This tutorial aims to explore the applications of LLMs in the context of medical text analytics.

The tutorial is designed to provide the audience with an introduction to LLMs especially within the healthcare domain. With a strong emphasis on practical applications, the objectives extend to demonstrate comprehensive skills to proficiently apply and refine these models.


History of LLM

20 minutes

Introduction to the history and background of LLMs starting from Transformer

LLM Key Terminology

15 minutes

Exploration of techniques and terminology used in LLMs


15 minutes

Overview of LLaMA family models and medical LLaMA models


10 minutes

Introducing Phi and our medical version, MedPhi2

Hands on session #1 Prompting

30 minutes

Learn how to use MedPhi2 on Google Colab

Tea/Coffee Break

30 minutes

Hands on session #2 Instruction Tuning

40 minutes

Train your own medical Phi2 on Google Colab

Multimodal LLM

20 minutes

Current research direction on integrating medical image with LLMs

Tutorial Material

GitHub Slides

Tutorial organisers

Yunsoo Kim

Yunsoo Kim is a PhD student at the Institute of Health Informatics (IHI), University College London (UCL), where he is actively involved in research related to medical large language models. Before his PhD, he was a senior research scientist and worked as the NLP group leader at LG Chem.

Jinge Wu

Jinge is a PhD candidate at IHI, UCL, where she is experienced in medical language models for various clinical applications. Jinge has presented her work at international conferences and collaborates with healthcare professionals to apply her research in practical settings.

Honghan Wu

Professor Honghan Wu is a Professor of Health Informatics and AI at the University of Glasgow, and currently hold an honorary Associate Professor position with the IHI, UCL where he leads a Health Informatics Group called KnowLab. His research interest is the intersection of health informatics and large language models.