AI to Protect Identifiable Data

Although healthcare generates vast amounts of data year after year, most of the data isn’t available because of the need to protect identifiable patient information. With limited data access, AI models often aren’t as reliable in the real world, which limits how they can be used within healthcare.

To expand AI applications while still protecting patient data, the Department of Energy (DOE) has committed $1 million towards a one year collaborative research project titled PALISADE-X: Privacy Preserving Analysis and Learning in Secure and Distributed Enclaves and Exascale Systems https://palisadex.net. The goal for the project is to create a secure AI framework to enable healthcare organizations to improve AI models used in biomedicine while keeping sensitive data secure.

DOE’s Argonne National Laboratory is leading the project in collaboration with DOE’s Lawrence Livermore National Laboratory (LLNL), University of Chicago, Broad Institute, and Mass General Brigham. Researchers will also collaborate with NIH to create a framework that supports Bridge2AI, a NIH program that’s developing new data sets to be used with AI to improve healthcare.

Within biomedicine, data used to train models can include information that is considered protected according to HIPAA regulations. To preserve privacy, organizations have avoided sharing their AI models or the data used to train them, and instead train their models using the limited data available to them internally.

However, using this approach, means that organizations are at risk of creating models that have bias. Since AI models can carry a lot of bias, they are not very effective in real world situations.

PALISADE-X will deliver a framework that can allow organizations to train AI models using data across multiple organizations, while keeping protected data secure. The framework will be developed as a software package for processing and securing data with advanced algorithms for federated learning, a form of machine learning that enables multiple organizations to collaboratively train a single model.

Once built, the PALISADE-X team will test the system by using AI models that predict the severity of COVID-19, using public and private biomedical datasets. Researchers also plan to use the framework to predict the risk of developing cardiovascular diseases.

If all is successful, the research will make it possible for organizations to develop and confidently share their AI models with scientists and relevant research groups without the worry of leaking private information.

According to Argonne Computer Scientist Ravi Madduri, “When we can safely share more data, we will be able to create better models that have less bias. When the tools are put in the hands of care provider, it can fundamentally change how medicine is practiced.”

This project is sponsored by the Office of Advanced Scientific Computing Research within DOE’s Office of Science. Go to https://science.osti.gov/ascr for more information

Share Button