Tamara Broderick first set foot on the MIT campus as a highschool student, as a participant within the inaugural event Women's Technology Program. The month-long summer academy offers young women a practical introduction to engineering and computer science.
What are the possibilities that she would return to MIT years later, this time as a college member?
Broderick could probably answer this query quantitatively using Bayesian inference, a statistical probability approach that attempts to quantify uncertainty by continually updating its assumptions as recent data becomes available.
In her lab at MIT, the newly appointed associate professor within the Department of Electrical Engineering and Computer Science (EECS) uses Bayesian inference to quantify uncertainty and measure the robustness of knowledge evaluation techniques.
“I actually have at all times been all for understanding not only what we all know from data evaluation, but additionally how well we understand it,” says Broderick, who can also be a member of the Information and Decision Systems Laboratory on the Institute for Data, Systems and society. “The reality is that we live in a loud world and can’t at all times get the precise data we would like. How can we learn from data while recognizing that there are limitations and coping with them appropriately?”
Broadly speaking, her focus is on helping people understand the restrictions of the statistical tools available to them and sometimes working with them to develop higher tools for a specific situation.
For example, her group recently worked with oceanographers to develop a machine learning model that could make more accurate predictions about ocean currents. In one other project, she and others worked with degenerative disease specialists on a tool that helps severely motor-impaired people use a pc's graphical user interface with the flick of a single switch.
A typical thread that runs through her work is the emphasis on collaboration.
“When you’re employed in data evaluation, you possibly can form of hand around in anyone’s backyard. It really can’t get boring because you possibly can at all times study one other subject and take into consideration how we will apply machine learning there,” she says.
Hanging out in lots of academic “backyards” is especially appealing to Broderick, as she had difficulty narrowing down her interests from a young age.
A mathematical way of pondering
Broderick grew up in a suburb of Cleveland, Ohio, and has been all for math for so long as she will remember. She remembers being fascinated by the thought of what would occur if you happen to continually added a number to itself, starting with 1+1=2 after which 2+2=4.
“I used to be possibly five years old, so I didn’t know what powers of two were or anything like that. I used to be just really all for mathematics,” she says.
Recognizing her interest in the topic, her father enrolled her in a Johns Hopkins program called the Center for Talented Youth, which gave Broderick the chance to take three-week summer courses on topics starting from astronomy to number theory to computer science.
Later, while at school, she conducted research in astrophysics with a postdoctoral fellow at Case Western University. In the summer of 2002, she spent 4 weeks at MIT as a member of the inaugural class of the Women's Technology Program.
She particularly enjoyed the liberty this system offered and its deal with using intuition and ingenuity to realize high-level goals. For example, the cohort was tasked with constructing a tool using LEGOs that allowed them to biopsy a grape suspended in Jell-O.
The program showed her how much creativity there may be in engineering and computer science and sparked her interest in an educational profession.
“But after I got to varsity at Princeton, I couldn't determine – math, physics, computer science – all of them seemed super cool. I desired to do all the things,” she says.
She selected to major in mathematics, but took all of the physics and computer science courses she could fit into her schedule.
Dive into data evaluation
After receiving a Marshall Scholarship, Broderick spent two years on the University of Cambridge within the United Kingdom, earning a Master of Advanced Study in Mathematics and a Master of Philosophy in Physics.
In the UK, she took quite a lot of statistics and data evaluation courses, including her first course on Bayesian data evaluation in machine learning.
It was a transformative experience, she remembers.
“During my time within the UK, I spotted that I actually enjoy solving real-world problems that matter to people, and that Bayesian inference has been utilized in a few of crucial problems of all,” she says.
Back within the United States, Broderick went to the University of California at Berkeley, where she joined Professor Michael I. Jordan's laboratory as a graduate student. She received her PhD in statistics with a deal with Bayesian data evaluation.
She selected a profession in academia and was drawn to MIT by the collaborative nature of the EECS department and the fervour and kindness of her prospective colleagues.
Her first impressions were confirmed, and Broderick says she found a community at MIT that helps her be creative and explore difficult, high-impact problems with far-reaching applications.
“I used to be lucky enough to work with a very great group of scholars and postdocs in my lab – sensible and hard-working individuals with their hearts in the precise place,” she says.
One of her team's most up-to-date projects involves working with an economist who’s studying using microcredit, the lending of small amounts of cash at very low rates of interest, in impoverished areas.
The aim of microcredit programs is to lift people out of poverty. Economists conduct randomized control trials with villages in a region that do or don’t receive microcredit. They wish to generalize the study results and predict the expected consequence of providing microcredit to other villages outside of their study.
But Broderick and her colleagues have found that the outcomes of some microcredit studies could be very fragile. Removing a number of data points from the info set can completely change the outcomes. One problem is that researchers often use empirical averages, where just a few very high or low data points can skew the outcomes.
Using machine learning, she and her collaborators developed a way to find out what number of data points to remove to alter the study's substantive conclusion. Using their tool, a scientist can see how brittle the outcomes are.
“Sometimes leaving out a really small portion of the info can change the important thing results of an information evaluation, after which we’d worry concerning the extent to which those conclusions generalize to recent scenarios.” Are there ways to make people aware of this? That’s what we’re trying to realize with this work,” she explains.
At the identical time, she continues to work with researchers in various fields reminiscent of genetics to know the benefits and drawbacks of assorted machine learning techniques and other data evaluation tools.
Happy trails
Research drives Broderick as a researcher, and it also fuels one in every of her passions outside of the lab. She and her husband enjoy collecting patches earned by climbing all the paths in a park or trail network.
“I feel my hobby really combines my interests of being outside and doing spreadsheets,” she says. “With these climbing areas you have got to explore all the things and then you definitely see areas that you just wouldn't normally see. It’s adventurous in that way.”
They've discovered some amazing hikes they never knew about, but they've also done greater than just a few “total disaster hikes,” she says. But each hike, whether hidden gem or overgrown mess, offers its own rewards.
And identical to along with her research, curiosity, open-mindedness and fervour for problem solving have never led her astray.
