|"Fault-tolerant and Distributed Data Analytics at the Edge" |
|Konuşmacı ||: ||Dr. Atakan Aral |
|Tarih ||: ||21 Ekim 2019 (Pazartesi) |
|Saat ||: ||14:30 |
|Yer ||: ||Bilgisayar ve Bilişim Fakültesi, |
İdris Yamantürk Konferans Salonu (1303)
The past decade has seen the rapid development of the IoT and the introduction of an entirely new generation of Internet services that radically changed many traditional industries. As a repercussion of this so-called smart revolution, there exists an ongoing paradigm shift from core data analytics to edge data analytics (e.g. Federated Learning). Edge computing is a natural fit as it enables processing with higher scalability, better privacy, and lower latency than centralized solutions. However, there does not exist a clear architecture for edge data analytics currently. In particular, we require novel solutions for fault-proneness and consistency challenges. In this talk, I will explain our approach to these challenges and proposed solutions over the course of three years within the scope of "Runtime Control in Multi Clouds" (RUCON) project at TU Wien. Some of the techniques that we utilized in these solutions include (dynamic) Bayesian networks and reinforcement learning.
Dr. Atakan Aral is a Research Fellow at the Institute of Information Systems Engineering, Vienna University of Technology (TU Wien). Prior to joining TU Wien in 2016, he had been a Research Assistant at the Faculty of Computer and Informatics Engineering, Istanbul Technical University (ITU) for five years. Dr. Aral received a dual MSc degree in Computer Science and Engineering from Politecnico di Milano (2011) and ITU (2012), and a PhD degree in Computer Engineering from ITU (2016). He participated in several research projects funded by organizations such as the European Research Council (ERC), Scientific and Technological Research Council of Turkey (TUBITAK), and Austrian Science Fund (FWF). Dr. Aral served on the Organizing Committee of IFIP/IEEE NOMS 2016 and as a TPC member in several leading conferences in his area including IEEE SERVICES and IFIP/IEEE IM. He was awarded six scholarships from institutions including Italian and Turkish governments for his academic success and is a co-author of the Best Student Paper at CLOSER 2018 as well as a Best Paper Candidate at ICFC 2019. His research interests center around resource management for distributed and virtualized computing architectures, failure resilience techniques for unreliable Edge resources, and distributed/federated machine learning systems that exploit Edge computing.