Solving Hard Engineering Problems with Deep Learning
Date/Time: September 24, 3:40 pm
Location: B29 Atanasoff Hall
Over the last few years, Deep Learning models have outperformed all other state-of-the-art machine learning techniques for many detection, classification and prediction problems. With the capability of handling very large data sets, these very large models extract hierarchical features at different scales of data without the need of hand-crafting. While at the lower layers, more primitive features are extracted, composite features (features of features) are extracted at the upper layers that are more expressive and often semantic in nature. However, most of the studied applications have been in the domains of computer vision, speech processing and natural language processing. This talk will discuss a couple of recent success stories of Deep Learning (i.e., unsupervised and supervised models, deep belief nets and deep convolutional nets) for two engineering applications that are rather nontraditional in the context of machine learning. The first one involves designing microfluidic lab-on-chip devices in real-time to achieve user-defined flow shapes that has enormous potential applications ranging from manufacturing new classes of polymerized fibers with engineered interactions to detecting human diseases. The second application is about early detection of flame instability in combustion processes via deep feature extraction from hi-speed flame images in order to prevent catastrophic lean blow out in aircraft and other engines.
Dr. Soumik Sarkar received his B. Eng. Degree in Mechanical Engineering in 2006 from JadavpurUniversity, Kolkata, India. He received M.S. in Mechanical Engineering and M.A. in Mathematics in 2009 from Penn State University. Dr. Sarkar received his Ph.D. in Mechanical Engineering from Penn State in 2011. He joined the Department of Mechanical Engineering at Iowa State as an Assistant Professor in Fall 2014. Previously, he was with the Decision Support & Machine Intelligence group at the United Technologies Research Center for 3 years as a Senior Scientist. Dr. Sarkar’s research interests include Statistical Signal Processing, Machine Learning, Sensor Fusion, Large volume information visualization, Fault Diagnostics & Prognostics, Distributed Control and Complexity Analysis withapplications to complex cyber-physical systems, robotics, thermo-fluid sciences and plant science. He co‐authored 53 peer-reviewed publications including 19 journal papers, 2 book chapters and one magazine article. Dr. Sarkar is currently serving as an Associate Editor of Frontiers in Robotics and AI: Sensor Fusion and Machine Perception journal.