“All of the breakthroughs that look like science fiction to us—artificial intelligence, self-driving cars, automated assistants, natural language recognition—are powered by data,” says assistant professor Theo Rekatsinas, who joined the Computer Sciences Department at the University of Wisconsin-Madison in fall 2017.
“While it might seem boring to manage data, the outcomes can be groundbreaking,” says the newest member of the department’s renowned database research group.
Behind-the-scenes work with data helps make today's technical advancements happen. For example, data integration entails working with data from wide-ranging sources that may vary widely in accuracy, format, how up-to-date the data is, and other factors.
Rekatsinas sums up the goal of data integration with a simple question: “How do you combine information from disparate sources to get insights?”
Skillful data integration can yield results with major impact. For example, Rekatsinas has worked on teams tackling important challenges like predicting civil unrest or disease outbreaks. By combing through various data points, including publicly available information like news articles, social media feeds and weather information, data scientists can predict phenomena with surprising accuracy. This work was funded by the Open Source Indicators program of the Intelligence Advanced Research Projects Agency (IARPA), a U.S. government agency.
Currently, he focuses on fundamental problems in the areas of data integration and data cleaning. "I'm developing algorithms and techniques that help us understand where the inherent noise in data comes from and removes it, leading to more accurate machine learning models," he says.
Rekatsinas was drawn to Madison on the strength of its database group, the first of its kind at an American university. UW-Madison professors were quick to see that this would become a major subfield of computer science.
“Madison is a powerhouse when it comes to databases; it's one of the best in the country and worldwide. It feels great to be a part of this team,” he says. Current colleagues in the database group are AnHai Doan, Jignesh Patel and Paris Koutris. Shivaram Venkataraman, who works on big data systems, will join the faculty in fall 2018.
A native of Crete, Greece, Rekatsinas completed his bachelor’s degree in electrical and computing engineering at the University of Athens. He then attended the University of Maryland for graduate school, followed by a two-year postdoc at Stanford University.
After trying out the East and West Coasts, he's now settling into life in Madison. He and his girlfriend foster rescue dogs for Fetch Wisconsin, giving pups a loving home and training them as they wait for adoption. Their other hobbies include hiking, cooking and concerts.
The academic lifestyle agrees with Rekatsinas. Seeing students' "aha moments" is what attracted him to university life rather than private industry, he says. “I like teaching to be extremely interactive, and I believe I give students these opportunities. It’s not the lecture style of Greece, like I had. The greatest thing is when they come to office hours as a group, and they begin to help each other."
Story courtesy of the Department of Computer Sciences.