AI Tool Predicts Disease Outbreaks Using Large Language Models

A groundbreaking AI tool, leveraging advanced language models, has been created to forecast the risk of infectious diseases. This novel approach, developed by researchers at Johns Hopkins and Duke Universities, surpasses current forecasting methods, potentially revolutionizing how public health officials manage outbreaks.

Revolutionary AI Framework Unveiled

Researchers have developed a new AI tool that utilizes large language modeling to predict the spread of infectious diseases. This tool, called PandemicLLM, reframes disease spread prediction as a text-based reasoning problem. It integrates complex, real-time, non-numerical information, outperforming existing models.

The team employed artificial-human collaborative design and time-series representation learning. The framework encodes multi-modal data for large language models, focusing on state-level spatial data, epidemiological time-series data, textual health policies, and genomic surveillance data.

“A pressing challenge in disease prediction is trying to figure out what drives surges in infections and hospitalizations, and to build these new information streams into the modeling,” commented co-corresponding author Lauren Gardner.

How It Works

The PandemicLLM model is trained on four key data types: state-level spatial data, epidemiological time-series data, textual health policies, and genomic surveillance data. By analyzing this information, the model predicts the interconnections and impacts on disease behavior.

“Traditionally, we use the past to predict the future. But that doesn’t give the model sufficient information to understand and predict what’s happening. Instead, this framework uses new types of real-time information,” said co-corresponding author Hao (Frank) Yang.

Testing the Model

To test the model, researchers applied it to the COVID-19 pandemic. The model was tested across all U.S. states for 19 months. The results showed that PandemicLLM outperformed existing models. It can also be adapted to forecast other diseases like bird flu, RSV, and monkeypox.

Currently, researchers are investigating how large language models can replicate human decision-making in health matters. The goal is to aid public health officials in formulating safer and more effective policies.

According to the CDC, the U.S. saw a 20% increase in flu-related hospitalizations during the 2022-2023 season, underscoring the need for improved predictive tools (CDC 2023).

Looking Ahead

Gardner concluded: “We know from COVID-19 that we need better tools so that we can inform more effective policies. There will be another pandemic, and these types of frameworks will be crucial for supporting public health response.”