Robert Chew named Bureau of Labor Statistics Senior Research Fellow

Chew will study the risks posed by sharing autocoding predictions externally
 

RESEARCH TRIANGLE PARK, N.C. — RTI International, a nonprofit research institute, today announced that senior research data scientist and program manager Robert Chew has received a Bureau of Labor Statistics (BLS) Data Science Fellowship.

As part of the fellowship, Chew will be asked to tackle the challenges of disclosure and privacy risks posed when publicly releasing autocoding models related to BLS’ Survey of Occupational Injuries and Illnesses (SOII).

“I am honored to be named a BLS Senior Research Fellow,” said Chew. “Conducting this research is a rare opportunity to benefit not only the academic community but also the broader U.S. public through helping BLS advance its mission.”

BLS is a leader in using artificial intelligence to automatically assign free-form text responses to standardized categories, such as occupations.  This process, also known as autocoding, helps improve the efficiency and reduce the time it takes for statistical agencies like BLS to generate official statistics. However, BLS is currently unable to fulfill requests from external parties interested in using its autocoders because of the possibility of unintentionally revealing sensitive or private information in the data used to train the model.

Chew will seek to better characterize the disclosure risk posed by sharing autocoding predictions with external parties. He will also work to develop privacy-preserving versions of the SOII autocoders to help reduce potential disclosure risk.

Chew holds a Master of Science degree in analytics from North Carolina State University’s Institute for Advanced Analytics and a Bachelor of Arts degree in economics and environmental studies from Oberlin College.

The BLS Senior Research Fellow Program is administered by the American Statistical Association under a grant from the National Science Foundation. It aims to encourage collaboration between academic scholars and government researchers in survey methodology, statistics, economics, data science and the social sciences.