Kira Goldner: Data Science and Computing for Social Good

Kira Goldner is a Shibulal Family Career Development Assistant Professor in the Faculty of Computing & Data Sciences at Boston University with a secondary appointment in Computer Science. As part of her role, she conducts research in the area of mechanism design for social good and soon, she is looking forward to advising students in research.

Prior to joining BU, Goldner completed her Bachelor of Arts at Oberlin College where she majored in Mathematics. Following her graduation, Goldner found that she was deeply interested in theoretical computer science and algorithms and went on to attend the University of Washington where she earned a Masters and PhD in Computer Science and Engineering. 

Goldner was drawn to the challenge presented by strategic interactions with algorithms, specifically mechanism design and the goal of revenue maximization as mathematically, this is one of the hardest challenges. Her dissertation was titled ‘ Mechanism Design for a Complex World: Rethinking Standard Assumptions’. She was also an NSF Mathematical Sciences Postdoctoral Research Fellow and a Data Science Institute Postdoctoral Fellow at Columbia University hosted by Tim Roughgarden.

However, after listening to a talk by Cynthia Dwork about differential privacy and fairness in machine learning Goldner’s interests changed.

“It sort of clicked – why am I doing revenue maximization when I could be doing something for society,” Goldner shares. “I still wanted to be doing the same fascinating math, but for society instead. So that became my goal: to use the same tools, but to help society.”

For Goldner, her main professional interests are participating in the community of people who are also interested in this intersection of incentives and algorithms, the EconCS community, as well as forming a community for individuals who are interested in mechanism design for social good.

At the moment, Goldner is working on expanding theory aimed toward social impact. 

“Foundational theory has a lot of assumptions in order to get to the most elegant solutions,” Goldner clarifies. “But that’s not always so realistic in practice. In order to make the theory apply to the world we live in, we need to expand the theory to assumptions that make sense.”

One model that Goldner is particularly interested in is called the interdependent values model – which was part of the 2020 Nobel Prize for Economics. To explain this theory, Goldner uses the example of the house-buying process.

“It basically encapsulates the fact that when we’re interested in buying a house, we’re not actually sure how much we’re willing to pay for the house because we don’t have all the information about the house,” Goldner says. “We don’t know how good the foundation is, how good the pipes in the bathroom are, or how good the roofing is, but maybe some other people know that information. So, our valuation depends on that information that other people have.”

Most prior theory assumes that once the individual pieces of information are learned, it is already publicly known how each individual uses that information to form their valuation. However, while this may be true in some situations, it is not always realistic in every scenario. Goldner is investigating this with coauthors Alon Eden and Shuran Zheng.

“The thing is, you don’t actually know how I combine information to form a valuation of how much I’m willing to pay for the house,” Goldner explains. “In settings where this is private, to know how much I’m willing to pay for the house, you would need to ask me how I came up with the valuation. Dealing with this private information and eliciting this private data – which is all very complex private information, requires a whole different approach and it’s very complicated.”

However, Golder asserts that individuals interested in data science should not be deterred by their lack of background knowledge or data science’s perceived complications. This is a view that she carries into the classroom when teaching.

“I don’t think it should matter the background that people have,” Goldner states. “I think the most important thing is people’s willingness to learn and the ability to teach oneself – if somebody is willing to come and learn no matter what it takes, then that’s what’s important.”

For those considering entering the data science and computing field but are unsure of where they stand, Goldner understands. When she first transitioned into computer science, Goldner felt that her database of knowledge was not on par with her peers and had to teach herself extra material in order to catch up. 

“Data science is exactly the area that says: it doesn’t matter what your background is so far, and it doesn’t matter where you’re going with it,” Goldner says. “It doesn’t matter what your before and after path is and how they match up with everyone else, they don’t have to be the same. This is where you come to learn the skills in between. This is where people match up in the middle.”