Postdoctoral research fellow Brittany Johnson wants to make engineered software more equitable through her research at the University of Massachusetts Amherst.
Johnson joined the College of Information and Computer Science’s (CICS) Center for Data Science (CDS) to improve software developer productivity. She works at the Laboratory for Advanced Software Engineering Research (LASER) with Yuriy Brun and Alexandra Meliou to develop new testing disciplines that help developers improve the quality and fairness of software.
Fittingly, her position is funded by the center’s Industry Affiliates Program (IAP), sponsored by the very businesses her research aims to help. IAP members pay a yearly fee for access to CDS research, recruiting, workshops and branding. Because her salary is funded by the IAP, she envisions a “stronger and more direct collaboration” between the business community and UMass computer and data science.
“I’m glad the CDS is shining a light on the importance of that collaboration by using that money to hire researchers like me,” Johnson said.
She said the disconnect between research progress and industry needs can slow productivity. She hopes her work will better connect researchers with developers and lead to the design of high quality software.
“Ideally, eventually you meet in the middle somewhere and come up with this awesome product that makes both research and industry better,” she said.
She is currently working with Brun and Meliou on a “fairness testing project,” investigating certain software for discrimination, especially racial and gender biases. The team at LASER developed Themis, which tests for and characterizes discrimination in software. Johnson explained that machine learning software can reflect the conscious or unconscious biases of the engineers or the data, so it is important to detect these biases and provide developers with the knowledge and tools to correct for them.
“It’s crazy how pervasive that concept is,” Johnson said. “We all have our notions and preconceptions, and sometimes they show up in software.”
Johnson is passionate that her work should make software easy to use and accessible to everyone. “It’s looking at fairness in software and being able to make software usable and approachable to everyone,” Johnson said. “It’s the idea of building tools that are useful for everyone no matter what level of understanding you have, no matter your experience or cultural or ethnic background.”
Beyond her research, Johnson wants to mentor other women and minorities in the field. She said, without the mentorship of her research advisor Jim Bowring during her undergraduate years at the College of Charleston, she would not have considered a Ph.D. Now, she wants to “make it full circle.”
“I would like to help other young ladies and other African American women and men, and make other minorities realize that being a minority doesn’t make you less qualified,” Johnson said. “We need that diversity. We need more females, more African Americans, more Latinos. I want that to be a part of my mission.”
One reason she wanted to pursue postdoctoral research was to mentor students. She enjoys engaging with students in the lab, and eventually anticipates a career in academia. She chose UMass because of its strong reputation in the machine learning research community and the high regard for Brun and LASER.
“Coming here was a huge career move, not just for me as a computer scientist, but working with Yuriy specifically as a software engineering researcher,” Johnson said.
As for UMass preparing her for academia, she said: “I want to be a Yuriy one day.”
Johnson is beginning to feel at home in the CDS community by speaking at, and attending events, such as Data Science Teas. She is looking forward to working with the local industry affiliates to make meaningful research contributions to further her career and provide long-term benefit to the Center for Data Science.
By Morgan Hughes