Enhancing Human-Computer Collaboration in Design Activities
Yi Ren, Department of Mechanical Engineering, University of Michigan
Tuesday, February 25, 2014
Goldwater Center (GWC) 487 [map]
Every day we observe the use of “crowd intelligence” in online services, e.g., Wikipedia, Innocentive and Amazon’s Mechanical Turk. How can a large crowd be engaged to help solving design problems? Can a machine (computer program) improve how it supports our design efforts by learning from a crowd? In this talk Ren will discuss how we utilize human knowledge to teach a machine what, for example, a good-looking car should look like for different consumers; and how design solutions could be crowdsourced, for example, for a hybrid powertrain architecture. The core research challenges include real-time adaptive interaction mechanism to improve the efficiency of learning from the crowd, and the learning of design heuristics from crowdsourced solutions. Challenges in handling noisy responses from anonymous crowd will also be discussed.
Yi (Max) Ren is a postdoctoral fellow at the Optimal Design Lab, University of Michigan, where he graduated in 2012 with a Ph.D. in mechanical engineering. His thesis work has led to a three-year NSF grant on collaborative human-machine interactions for creative design using crowdsourcing. His research interests include design optimization, machine learning and configuration design.