Silvijia Kokalj-Filipovic of Rowan University

Abstract: The first part of my talk will describe some of my research in the domain of machine learning on radio-frequency signals (RFML) – in particular deep learning for RF applications used in wireless communication and spectrum sensing. The second part will shed light on a new area of AI, popularized as learned compression, which in many data-domains achieved better results than the best traditional lossy information compression methods. I will illustrate it using my recent research results

Biography:

Dr Silvija Kokalj-Filipovic is a Professor of Computer Science at Rowan University. Her career combines the fields of wireless communication networks, signal processing and artificial intelligence, and the experiences of a cellular technology developer where she got inspired to do PhD in wireless communications. She later worked as a scientific researcher at industrial, academic and government research labs. Her research career resulted in several grants from the DoD, including DARPA and ONR. As a PI she worked on applying the latest AI on the modern wireless technology, including implementation of neuromorphic learning algorithms and designing distributed machine learning over wireless networks. She also conducted pioneering research on the topic of adversarial attacks on AI models used in radio-frequency machine learning (RFML). She made contributions in the RFML-based spectrum awareness and complex wireless systems, and was active in building the RFML research community by serving as the Lead Guest Editor for the IEEE Journal of Selected Topics in Signal Processing (JSTSP) - special issue on ’Machine Learning for Cognition in Radio Communications and Radar’, 2018, and as chair of multiple workshops on Machine Learning for Wireless Communications.