The Federal Chief Data Officers (CDO) Council is looking for information on synthetic data generation as it works to develop a best practice guide, according to a request for information (RFI).
The General Services Administration posted the RFI on the Federal Register on Jan. 5, noting that it wants input from the public, private, and academic sectors.
“The CDO Council is interested in consolidating feedback and inputs from qualified experts to gain additional insight and assist with establishing a best practice guide around synthetic data generation,” the RFI says. “The CDO Council has preliminarily drafted a working definition of synthetic data generation and several key questions to better inform its work.”
The National Institute of Standards and Technology (NIST) defines synthetic data generation as “a process in which seed data is used to create artificial data that has some of the statistical characteristics as the seed data.”
The council said that it believes this definition also includes techniques such as privacy-preserving synthetic data generation and generative adversarial networks (GANs), which use a deep learning model to generate synthetic data that mimics real data.
The CDO Council noted that the Federal government would benefit from developing a more formalized definition for synthetic data generation, and it wants to know if there are any limitations to relying on the NIST definition.
Aside from defining synthetic data generation, the council also wants to hear examples of synthetic data generation, challenges and limitations to consider, and best practices for CDOs.
Comments are due by Feb. 5, 2024.
