Walter Lasecki, a graduate student at the University of Rochester, has been awarded the 2013 Ph.D. Fellowship at Microsoft Research. The fellowship is a highly competitive award given to 12 outstanding Ph.D. students in North American institutions each year in support of their academic research.
"Walter's award reflects both his brilliance and our department's success in graduate education," said Henry Kautz, chair of the computer science department at Rochester. "It is great to see University of Rochester alongside of Berkeley, CMU, MIT, Johns Hopkins, and Penn in the list of the winners' universities."
It is no surprise that Lasecki, who started programming while in elementary school, has focused on computer science. His projects aim to create a collective intelligence capable of completing tasks better and more consistently than any individual member of the crowd using mediation by an automated process.
"We're thinking about how to have a conversation with the crowd," said Lasecki. "There's gap between human computation and machine computation: they're not speaking the same language." He explains that although a phrase expressed in natural language – that is the way people normally speak – might be perfectly understandable to humans, a machine doesn't know what to do with it. Lasecki investigates how to connect the two using human understanding.
Now a third-year Ph.D. student, Lasecki has been working with his adviser Professor Jeff Bigham on projects that leverage this interaction to help people with disabilities. Together, Lasecki and Bigham have developed "Legion:Scribe," which allows real-time captioning of audio for the hard-of-hearing.
"It really lowers the barriers to entry; instead of paying a professional stenographer $150 an hour scheduled in advance, you can have access to a group of people at a fraction of the cost and on demand," Lasecki said. "We have very high precision and can send text back within a few seconds hearing it."
This builds on previous work in assistive technology by Bigham. An example of another project is a system that allows the visually-impaired to take a photo of their surroundings and quickly get feedback on what the crowd sees. For example, a user could hold a can of food up to a camera and get feedback from the crowd, telling him what the label on the can says.
"Walter is a truly rare graduate student who enters the program with an idea of what he wants to work on and brings people together around that idea to make it happen," said Bigham.
Lasecki's work has been previously covered in the New York Times, the MIT Technology Review, Gizmodo, and other publications.
The award covers the tuition and fees for two years of graduate study, a stipend for living expenses, and the opportunity to complete a salaried internship at Microsoft during the year following the award. Fellowship winners this year include students from Stanford University, the Massachusetts Institute of Technology, and Carnegie Mellon University.