Friday, 27 January at 3pm
Math & Science Center, Room W301
By combining machine and crowd intelligence, we can open up a broad new class of software systems that solve problems neither approach could solve alone. However, while crowds are increasingly adept at straightforward parallel tasks, they struggle with complex work because participants vary in quality, well-intentioned contributions can introduce errors, and future participants amplify and propagate those errors. I introduce techniques that decompose complex tasks into simpler, verifiable steps, and return crowd results in realtime. I use these techniques to create crowd-powered systems: interactive applications that react with a combination of human and algorithmic intelligence. These systems support goals like rewriting and shortening text, polling opinions within seconds, and taking better photographs.
Michael Bernstein is a PhD candidate in computer science at the Massachusetts Institute of Technology. His research in human-computer interaction focuses on crowdsourcing and social computing systems. He has been awarded Best Student Paper at UIST 2010, Best Paper at ICWSM 2011, the NSF graduate research fellowship and the Microsoft Research Ph.D. fellowship. His work has appeared in venues like the New York Times, Slate, CNN and Wired. He earned an S.M. in Electrical Engineering and Computer Science from MIT, and a B.S. in Symbolic Systems from Stanford University.