Wednesday, September 30, 2020 | 12pm to 12:45pm
About this Event
Preseneter: Assistant Professor Faez Ahmed, MechE
Abstract:
Modern machine learning techniques, such as deep neural networks, are transforming many disciplines ranging from transportation to healthcare, by uncovering patterns in big data and making accurate predictions. They have also shown promising results for discovering high-quality novel design ideas, which is crucial for designing products and enabling innovation. These automated computational design methods can support human experts, who typically create designs by iteratively exploring ideas using experience and heuristics, that can be time-consuming or inadequate for exploration. In this talk, we will discuss how design problems present their unique challenges to computational methods and how some of our recent work address these challenges. Specifically, we will discuss how a new type of Generative Adversarial Network-based method, which can learn compact representations, outperforms state-of-art design parametrization methods with more than 186% improvement for the design optimization of airfoils. We will also discuss broader applications and open challenges in data-driven design.
Bio:
Dr. Faez Ahmed is an Assistant Professor in Mechanical Engineering at MIT, where he leads the Design Computation and Digital Engineering (DeCoDE) lab. His research centers on studying complex engineering design problems by developing new combinatorial optimization and machine learning methods. Before joining MIT, Faez was a Postdoctoral Fellow at Northwestern University and did his Ph.D. in Mechanical Engineering at the University of Maryland. He also worked in the railway industry in Australia, where he led efforts on developing data-driven predictive maintenance methods. His recent work has focused on creating the first provably-optimal algorithm for the diverse matching problem, proposing AI-driven synthesis methods to generate novel designs, and building computationally efficient ways for designing engineering material systems.
Zoom information: https://mit.zoom.us/j/95791474660