CS Seminar: “Learning for Intelligent Systems,” Dr. Sedat Özer (University of Central Florida), EA-409, 1:40PM March 11 (EN)

S E M I N A R: Learning for Intelligent Systems

Dr. Sedat Özer
University of Central Florida, USA

Recent developments in deep learning field put neural networks at the core of intelligent systems. In today’s connected world, self-driving vehicles, cell phones, laptops and smart devices stream data in a never-ending way and as a collective data source, they form ideal environment for deep learning based intelligent systems to learn. In this talk, I focus on learning from a particular data format: images and videos. In the first part of my talk, I introduce my novel learning algorithm that is related to a particular neural network type: Radial Basis Networks. My algorithm can visualize its parameters on 2D binary shapes intuitively. I also demonstrate how that property of the algorithm can be used for parametric shape modeling. In the second part of my talk, I present two different and recent projects that I have been working on utilizing deep learning. For the first application, I present my ongoing work on anomaly detection in surveillance videos. As for the second application, I present my work on controlling the steering mechanism in self-driving vehicles in cooperative environments where the vehicles communicate with each other.

Bio: Dr. Sedat Ozer received his Ph.D. degree from Rutgers University, 2013, where he worked on analysis and visualization of time-varying volumetric data sets. As a research associate, he worked at multiple research institutions including Virginia Image and Video Analysis (VIVA) Lab at the University of Virginia, Computer Science and Artificial Intelligence Lab (CSAIL) at the Massachusetts Institute of Technology (MIT) and Center for Research in Computer Vision (CRCV) at the University of Central Florida. His research interests cover developing new, scalable and explainable machine learning algorithms, developing techniques for data reduction, data fusion and data analysis applications for intelligent and self-driving systems. He serves as a publicity co-chair for IEEE Connected and Automated Vehicles Symposium in 2019. He also serves as a reviewer in multiple IEEE transactions and conferences related to computer vision, pattern recognition, neural networks and image processing.

DATE: 11 March 2019, Monday @ 13:40