I recently met a remarkable young woman, the child of immigrants from southern India, who has yet to enter graduate school but has already completed fundamental research on the nature of artificial intelligence.
Her name is Pratyusha “Priya” Kalluri, and she’s from America’s heartland, Madison, Wisconsin—though when I spoke with her she was in Spain doing computer research at the Complutense University of Madrid. In the fall, she’ll be entering the graduate program at Stanford University.
Her family’s emphasis on education motivated Pratyusha to pursue an undergraduate degree at MIT.
In an early project, she built systems to reveal the goings-on inside the human body: at the Dana-Farber Cancer Institute, she developed an algorithm to identify the gene pathway changes that underlie breast cancer. In this work, she took the approach of many AI researchers—examining how to apply computer intelligence to existing practical endeavors, opening up new vistas into the human body. She created data mining software able to analyze large datasets about many patients in order to enable scientists to spot the key genetic changes that signal the onset of an aggressive cancer. In a phone call from Madrid, Priya told us: “If you run these algorithms they point to groups of genes that have functionally clustered. You can see that as cancer becomes increasingly fatal, these genes have become hyper correlated like they’ve been coopted by the cancer and are now in overdrive.” That knowledge can lead to treatments targeting those genes.
Her work created a new algorithm for processing large quantities of data, not only data related to cancer. Eventually, at MIT, this passion for building smart, useful computer systems led her to become enthralled by a passion to understand the most fundamental principles of artificial intelligence: how to create systems that behave in smarter and human-like ways. As an undergraduate at MIT and subsequently a visiting researcher in Madrid, Pratyusha built AI systems modeling facets of human intelligence and human language processing.
Remarkably, she has been trying to find the most fundamental software modules that will enable a computer to understand the social meaning of a given human behavior. For example, her work could eventually enable a computer system to understand the difference between anger and sadness, say, in any given scenario. This is incredibly difficult, partly because scientists still don’t fully understand how the human brain itself works, let alone how to duplicate what it does. It’s also complex because people naturally learn to read many subtle indicators that tell whether someone else is hostile or friendly.
She relied on research into how children learn—how they begin very early in recognizing hindering and helping behaviors—and she allowed this to guide her work. If these core modules grow and become robust enough, scientists could then link them to other modules to give greater and greater complexity and subtlety to an AI system—for example, enabling systems to detect and respond to deception. Something this commonplace for human beings would be an enormous achievement for computer engineers and far into the future.
Priya has recently won a Soros Fellowship, and it will become central to who she is able to become in the next few years, and maybe for the rest of her life. It will enable her to do research without worrying about funding. When looking for grants, from any source, to fund research projects, it’s easier to win them by proposing work that has a measurable practical goal; it can be much more difficult to find money for exploratory research that doesn’t get immediate results. This will give Priya the ability to pursue that harder-to-fund basic research.
All of this is possible because Paul and Daisy Soros wanted to do something for immigrants, like themselves. In 1997, they created The Paul and Daisy Soros Fellowships for New Americans. They wanted to find people who, like them, had the ability and desire to handle an advanced education, but needed the sort of financial help that wasn’t available to them when they came here. Field of study didn’t matter.
They had to simply show a likelihood of becoming leaders and contributors and also have an understanding and respect for the principles and values of this unique nation: freedom and individual rights. With an endowment of more than $90 million, the Fellowship selects 30 applicants for up to $90,000 to cover education costs and expenses over two years. From a pool of 1,700, only 30 are selected.
“When I have conversations with professors, and I say I have external funding from this Fellowship, there is suddenly added freedom. I can conduct research off-the-beaten-path, or I don’t have to stick to just one research path. Simultaneously, from the PD Soros application to the interviews to the Soros community I am joining, I am connecting to people excited about the beautiful power of immigrants and New Americans.”
Priya is joining a growing community of gifted immigrants who are giving back to American culture far more than they were given by the foundation—an investment in immigrants is an investment in America’s future.
Peter Georgescu is the author of Capitalists Arise.