Tamara Broderick first set foot on MIT’s campus as a high school student as a participant in the inaugural Women in Technology Program. The month-long summer academic experience provides young women with a hands-on introduction to engineering and computer science.
What are the chances that she will return to MIT in a few years, this time as a faculty member?
This is a question Broderick could probably answer quantitatively using Bayesian inference. Bayesian inference is a statistical approach to probability that attempts to quantify uncertainty by continually updating assumptions as new data is obtained.
In his lab at MIT, a newly tenured associate professor in the Department of Electrical Engineering and Computer Science (EECS) is using Bayesian inference to quantify uncertainty and measure the robustness of data analysis techniques. Masu.
“I’m always very interested in understanding not just ‘what can we learn from data analysis,’ but ‘how well do we know it?'” says Broderick, who is also a member of the institute. Research Institute for Data, Systems and Society. “The reality is that we live in a noisy world and we don’t always get exactly the data we need. Learn from the data, but at the same time recognize that there are limitations and address them appropriately. How can I do that?”
Broadly speaking, she focuses on helping people understand the limitations of the statistical tools available to them, and in some cases working with them to create better tools suited to specific situations. Sometimes.
For example, her group recently worked with oceanographers to develop machine learning models that can make more accurate predictions about ocean currents. In another project, they are collaborating with degenerative disease experts to develop a tool that allows people with severe motor disabilities to utilize a computer’s graphical user interface by operating a single switch. We worked on the development of
A common thread woven through her work is an emphasis on collaboration.
“When you work in data analysis, you get to be in and out of everyone’s backyard, so to speak. It’s really boring because you’re always learning about other fields and thinking about how machine learning can be applied there.” There’s nothing to do,” she says.
Wandering around the many academic “backyards” was particularly appealing to Broderick, who struggled to narrow down his interests from an early age.
mathematical thinking
Broderick, who grew up in the suburbs of Cleveland, Ohio, has been interested in mathematics ever since he could remember. She recalls that she started with 1 + 1 = 2 and was fascinated by the idea of what would happen if she kept adding the numbers to 2 + 2 = 4.
“I was probably 5 years old, so I didn’t know what a ‘power of two’ was or anything like that. I just loved math,” she says.
Her father recognized her interest in this field and enrolled her in a program at Johns Hopkins University called the Center for Talented Youth. This gave Broderick the opportunity to take a three-week summer course with her on a variety of topics, from astronomy to number theory to computer science.
Later, in high school, he conducted research in astrophysics with a postdoctoral fellow at Case Western University. In the summer of 2002, she spent her four weeks at MIT as a member of the inaugural class of the Women’s Technology Program.
She particularly enjoyed the freedom the program offered and its focus on using intuition and ingenuity to achieve high-level goals. For example, the research group was challenged to build a Lego-based device that could be used to biopsy grapes suspended in Jell-O.
The program showed her how much creativity is involved in engineering and computer science and piqued her interest in pursuing an academic career.
“But when I got to Princeton, math, physics, computer science, they all looked so cool, I couldn’t decide. I wanted to do them all,” she says.
She decided to pursue a bachelor’s degree in mathematics, but took all the physics and computer science courses she could squeeze into her schedule.
Digging into data analysis
After receiving the Marshall Scholarship, Broderick spent two years at the University of Cambridge in England, earning a Master of Advanced Studies in Mathematics and a Master of Philosophy of Physics.
In the UK, I took a number of classes in statistics and data analysis, including my first class on Bayesian data analysis in the field of machine learning.
It was a transformative experience, she recalls.
“During my time in the UK, I realized that I really enjoy solving real-world problems that matter to people, and that Bayesian inference is used in some of the most important problems,” she said. say.
Returning to the United States, Broderick headed to the University of California, Berkeley, where he joined the laboratory of Professor Michael I. Jordan as a graduate student. She earned her PhD in statistics with an emphasis on Bayesian data analysis.
She decided to pursue a career in academia and was drawn to MIT by the collaborative nature of the EECS department and the passion and friendliness of her future colleagues.
His first impressions were spot on, and Broderick says he found a community at MIT that helped him unleash his creativity and explore difficult and impactful problems with a wide range of applications.
“I’ve been lucky to work with some really great students and postdocs in my lab. They’re bright, hard-working people with their hearts in the right direction,” she says.
One of her team’s recent projects involves a collaboration with economists who are studying the use of microcredit, or the lending of small amounts of money at very low interest rates, in poor communities.
The goal of microcredit programs is to lift people out of poverty. Economists conduct randomized controlled trials in local villages that do or do not receive microcredit. They hope to generalize their findings and predict the expected results if microcredit is applied to other villages not included in the study.
But Broderick and his colleagues found that the results of some microcredit surveys can be very fragile. Removing one or a few data points from a dataset can completely change the results. One problem is that researchers often use empirical averages, and a small number of very high or low data points can skew the results.
She and her collaborators have developed a method that can use machine learning to determine how many data points need to be removed to change a study’s substantive conclusions. Their tools allow scientists to see how fragile their results are.
“Removing even a small portion of the data can change the key results of a data analysis, and in that case, you may be concerned about how well those conclusions generalize to new scenarios. Let people know. Is there a way? That’s what we’re trying to do with this study,” she explains.
At the same time, we continue to collaborate with researchers in various fields, such as genetics, to understand the strengths and weaknesses of different machine learning techniques and other data analysis tools.
happy trail
Exploration is what drives Broderick as a researcher, and it also fuels one of her passions outside the lab. She and her husband enjoy hiking all the trails in her system of parks and trails and collecting the patches they earn.
“I think my hobby is a combination of my interest in the outdoors and my interest in spreadsheets,” she says. “These hiking patches allow you to explore everything and see areas you wouldn’t normally see. It’s adventurous in that sense.”
Not only did they discover some amazing hikes they had never known about, she says, but they also took several “total disaster hikes.” But each hike has its own rewards, whether it’s a hidden gem or a grassy wasteland.
And just like her research, her curiosity, open-mindedness, and passion for problem-solving never led her astray.