Researchers are working on a computer model that anonymizes sensitive information in research data, hiding potentially identifiable information.
The project, led by Pacific Northwest National Laboratory (PNNL) Data Scientist Tony Chiang, DPhil, and University of Texas at El Paso (UTEP) Mathematical Sciences Professor Amy Wagler, Ph.D., aims to protect the privacy of individuals who contribute to studies without compromising the integrity of research data.
To do this, the researchers use machine learning models to transform “real-world data into artificial data.”
Wagler said the computer model would be particularly useful in the health sector, where healthcare providers are cautious of breaching patient confidentiality by sharing private data.
How the Computer Model Protects Privacy
The computer model the researchers are working on will analyze original research data and use statistical models to generate artificial data that resembles it in every way.
“The artificial data will resemble the original data in every statistical aspect and characteristic but will be shareable without compromising the privacy of those who contributed the data — and researchers will still be able to gain valuable insight from it,” a blog post on the UTEP’s website explains.
A separate “adversary” model will analyze the two data sets and attempt to distinguish between them.
“As soon as the model cannot discriminate between the two sets of data, we know we can work with it,” Wagler said. She explained that the goal is to make it “highly improbable” that individuals can be identified from research data.
It’s unclear when the new computer model will be launched.
Research ethics requires researchers to protect the privacy of their subjects. However, it can be difficult to do this when dealing with large data sets. And even when researchers take precautionary steps to eliminate personally identifiable information from their research data, it may be possible to infer certain information about the subjects who provided the data.
Breaking Down ‘Academic Silos’ Through Collaboration
Chiang and Wagler maintain joint appointments with UTEP and PNNL. Wagler highlighted the importance of collaboration in the academic field.
“We are often siloed in academia, but working on shared projects with PNNL breaks down those silos and provides opportunities, professional support, and resources we need to solve these challenges,” Wagler noted.
Chiang emphasized the value of mentoring Ph.D. students at PNNL.
“The most rewarding part of the collaboration is getting to work with Ph.D. students and helping them open their eyes to see the impact of their work. “Often, Ph.D. students are solving problems for a thesis without a clear implication for the usefulness and utility of their work.”
Don’t miss out on important developments in the cybersecurity and data privacy field. Follow us on Twitter, Threads, and Mastodon for the latest news!
