Data-Driven Engineering Design: Design democratization in the age of machine learning

Design democratization — the ability of people all over the world to collaborate together to design physical products — can transform the way we traditionally think about designing products. However, to enable design democratization, we need to support it with machine learning and computing methods. In this talk, I will talk about three problems faced by organizations in gathering and processing ideas from distributed teams: 1) How does one form teams to evaluate design ideas? 2) How does one reliably measure the creativity of ideas? and 3) How does one filter good ideas out of hundreds of submissions? I will discuss how matching, ranking, and novelty estimation algorithms developed by me address parts of these problems. The scientific and mathematical work done to answer these questions lay the foundation to understand representation, learning, and optimization of discrete items and apply widely to many other fields.

Bio of speaker: Faez Ahmed is a Ph.D. candidate in Mechanical Engineering at the University of Maryland, College Park. He did his undergraduate and master at Indian Institute of Technology, Kanpur. Prior to his Ph.D., he worked as a Reliability Engineer for railways in Western Australia. He is a Future Faculty Fellow at the University of Maryland and recipient of the Kulkarni Fellowship. He is also the lead organizer of ACM New York non-profit, to spread computing in the local community. Faez works at the intersection of Machine Learning, Engineering Design, and Human-Computer Interaction to enable globally distributed teams of designers to participate in the design process.  His interests are in studying principled methods for representation, learning, and optimization of discrete problems occurring in design.