DHS Seeks Data Generation Platforms for Machine Learning Security, Privacy

The Department of Homeland Security’s Science and Technology Directorate is seeking vendors and suppliers of secure synthetic data generation platforms. 

S&T said synthetic data will enable training in DHS machine learning models when real-world data is unavailable or when their use would jeopardize privacy and security.

According to Melissa Oh, S&T Silicon Valley Innovation Program director, scaled synthetic data generation is essential in safeguarding data privacy, civil rights and liberties, the DHS said Friday.

S8T is seeking versatile data generation solutions that maintain data realism and are capable of supporting structured and unstructured data. The data generator should also be capable of replicating data sets with statistical attributes and generating quality data for ML models. In addition, S&T wants solutions that could prevent reverse engineering of synthetic data.

DHS operational components and offices, such as the Cybersecurity & Infrastructure Security Agency and the DHS Privacy Office, are the intended users of the synthetic data generators.