The Facebook Blog

Earlier this year, we launched the Facebook Fellowship Program to support Ph.D. students in the 2010-2011 school year who can help solve some of the biggest challenges facing the social web and Internet technology. We received many hundreds of applications from students at U.S. universities, and we were amazed with their talent and range of accomplishments.

We started this program because the academic community plays a central role in addressing many of our most challenging research questions on topics ranging from cloud and social computing to Internet economics and machine learning. Identifying the five winners listed below was a difficult process, and in the end we looked for students that had a history of academic accomplishment, represented a diversity of research fields, and were working on topics that are directly applicable to the social web and Facebook.

Our first class of fellows include students researching crowdsourced online help, the structure of networks, information disclosure, online advertising and data analysis. In addition to the five winners, we chose 22 finalists who each received $500.

We think all of the winners and finalists represent some of the brightest minds in academia today, and we're looking forward to their continued research. Each fellow receives paid tuition and fees, a $30,000 stipend, conference travel and other benefits.

Below are profiles of each of the Facebook Fellowship Program winners (in alphabetical order), along with insight in their own words about their areas of research. Congratulations to each of them—we're proud to call them our Facebook Fellows!


Vinayak Borkar, University of California at Irvine
Area of Focus: Cloud Computing

Vinayak is focusing his research on ways to improve distributed computing platforms for data analysis and applied for a fellowship because of the complex data processing challenges Facebook is tackling. Currently a visiting researcher and doctoral student at the University of California at Irvine, Vinayak previously spent 8 years as a software engineer at various software companies and most recently was the co-founder and chief architect of Black Titan Software in San Jose, Calif.

Quote: "Large-scale data processing is undergoing a radical change. Innovation in these areas is happening at places like Facebook. I look here for interesting data problems that will push the frontiers of research. On a prior visit to the Facebook office I realized that my research goals coincided greatly with some of the challenges faced by the Facebook developers. I applied for the fellowship to collaborate with the Facebook team, with the hope of solving some exciting real-world problems."


Parmit Chilana, University of Washington, Seattle
Area of Focus: Social Computing

Chilana

Parmit is pursuing research in human-computer interaction and applied for a fellowship in order to apply her ideas for crowdsourcing help for web applications in a real-world setting. Parmit is a PhD candidate in information science at The Information School at the University of Washington. She previously earned a master's degree in library and information science from the University of Illinois at Urbana-Champaign and a bachelor's degree in computer science from Simon Fraser University in British Columbia, Canada.

Quote: "To most people, Facebook is largely a means for staying in touch with friends and social networking. But, I want people to know how powerful this type of large-scale online participation, community-building, and wisdom of the crowds or crowdsourcing is for tackling some important problems in human-computer interaction today.

For example, in my current work, I'm developing techniques for supporting just-in time, contextual help for web applications where users can point to something within the interface and ask questions about resolving technical issues. What makes this approach unique is that the solutions will be 'crowdsourced' from communities of users. This will allow people to not only learn from each other's experiences, but also to get help more efficiently than using other means, such as searching separate, disconnected help resources."


Leslie John, Carnegie Mellon University, Pittsburgh
Area of Focus: Behavioral Economics

John

Leslie is researching new dimensions of privacy and how people decide what information to disclose. She applied for a fellowship to extend her research to the context of social networking. Leslie is a doctoral candidate in behavioral economics at Carnegie Mellon. She previously earned a master's degree in psychology and behavioral economics from Carnegie Mellon and a bachelor's degree in psychology, arts and business from the University of Waterloo in Canada.

Quote: "I've been doing research at the intersection of psychology, economics, and marketing; my primary focus is to try to understand when and why people are willing or unwilling to divulge personal information.

With collaborators George Loewenstein and Alessandro Acquisti, I've been finding that people's concern for privacy and propensity to self-disclose can be influenced, both upward and downward, by factors that are difficult to justify. For example, cues that should alert people to a threat of privacy can instead suppress privacy concern, and elicit self-disclosure. Likewise, we've also been finding that cues that should calm privacy concern can instead do the opposite, by rousing privacy concern and suppressing revelation."


Mladen Kolar, Carnegie Mellon University, Pittsburgh
Area of Focus: Machine Learning

Kolar

Mladen is studying the structure of networks and how they change and evolve over time. Mladen's investigations cover a wide range of networks—he has looked at how senators' positions and interactions have changed based on their voting records as well as reverse engineered how genes regulate each other over time. Mladen is a Ph.D. student in the machine learning department at Carnegie Mellon, and he is a graduate of University of Zagreb, of Zagreb, Croatia, in electrical engineering and computing.

Quote: "The main premise of machine learning is that the prediction accuracy should improve as more data become available. Unfortunately, achieving improved accuracy demands complex predictive models, which are often hard to interpret for the purpose of scientific discovery. In my research, I develop flexible statistical models, which can provide insight into complex patterns underlying high-dimensional and noisy data, and efficient computational algorithms for fitting these models.

The main theme is to trade off prediction accuracy for simpler models that help understand the complex nature of data. Much of my research has focused on models for networks, which are simple, yet powerful tools for capturing relationships among individuals. Networks help us answer some of the fundamental questions of interest, such as: 'What role(s) do individuals play when they interact with different peers?' and 'How do social groups form and dissolve as a response to external stimuli?' "


Yaron Singer, University of California at Berkeley
Area of Focus: Internet Economics

Singer

Yaron's research focuses on algorithmic game theory and mechanism design. His interest in Internet economics already has led him to found an advertising network for social media, Bidwave, and he plans further work into how to build a platform where people view ads as beneficial recommendations. Yaron is a Ph.D. candidate in computer science at the University of California at Berkeley. He previously graduated with a dual degree in mathematics and computer science from Tel-Aviv University in Israel.

Quote: "The emergence of online social networks takes Internet economics to new frontiers. With open application platforms and APIs and large-scale network data, social networks bring a new dimension to the web, forcing us to rethink standard solutions inherited from previous models. Researchers and engineers at Facebook are at the forefront of these efforts, constantly pushing innovation forward. The opportunity to be exposed to interesting new problems whose solution may have a significant impact, seems rare and fascinating."


Greg Badros, a University of Washington Ph.D. '00 and a director of engineering at Facebook, wishes his house understood him better before it talks back.