The review of the TEWI colloquium of Prof. Dr.-Ing. Jörn Ostermann from July 12, 2018 comprises the slides (below) and video here.

Abstract: We present a comparison between AV1, HEVC and JEM. It reveals that professionally optimized AV1 encoding software is about 32 times as complex as the HEVC encoder software HM. Given the wide attention that AV1 receives and the plans of Google and Facebook to actually use AV1 in commercial environments, the complexity constraints endured during the HEVC development seem to disappear. We present recent results on contour-based prediction for intra coding, a KLT depending on scene and intra prediction direction for coding the prediction error and texture synthesis to replace conventional texture coding for uniform parts of images. Especially contour-based prediction and texture synthesis depend heavily on computer vision algorithms. While typically failures of efficient prediction can be compensated by coding the prediction error and hence a higher data rate, texture synthesis in video coding does not have this safety leash.

Bio: Jörn Ostermann studied Electrical Engineering and Communications Engineering at the University of Hannover and Imperial College London, respectively. He received Dipl.-Ing. and Dr.-Ing. from the University of Hannover in 1988 and 1994, respectively. From 1988 till 1994, he worked as a Research Assistant at the Institut für Theoretische Nachrichtentechnik conducting research in low bit-rate and object-based analysis-synthesis video coding. In 1994 and 1995 he worked in the Visual Communications Research Department at AT&T Bell Labs on video coding. He was a member of Image Processing and Technology Research within AT&T Labs – Research from 1996 to 2003. Since 2003 he is Full Professor and Head of the Institut für Informationsverarbeitung at Leibniz Universität Hannover, Germany. Since 2008, he is the Chair of the Requirements Group of MPEG (ISO/IEC JTC1 SC29 WG11). Jörn was a scholar of the German National Foundation. In 1998, he received the AT&T Standards Recognition Award and the ISO award. He is a Fellow of the IEEE (class of 2005). Joern served as a Distinguished Lecturer of the IEEE CAS Society (2002/2003). He published more than 100 research papers and book chapters. He is coauthor of a graduate level text book on video communications. He holds more than 30 patents.

Alois Reitbauer |  August 23, 2018 | 15:00 | E.2.42

Abstract

You may have heard of autonomous self-driving cars, but what is autonomous self-driving software. Dynatrace has been running their software systems following a NoOps approach for several years now. Based on their experience they have developed a new approach to manage software applications using concepts like unbreakable delivery pipelines and self-healing deployments. Learn what is behind the idea of NoOps and how to build applications that run and manage themselves, what can be built today and what the future will bring.

CV

Alois Reitbauer is the Chief Technology Strategist of Dynatrace and leads the Dynatrace innovation lab. Alois has successfully developed several solutions in the application performance management space and brought them to market. Currently he is working with his team on building autonomous software systems that manage themselves. Alois is further interested in the impact of new collaboration technologies like voice or Augmented Reality. In both areas he currently works with customers to develop market-ready products. Alois is also a frequent speaker at technology conferences and the chair of the W3C Distributed Tracing working group.

Prof. Dr.-Ing. Jörn Ostermann | July 12, 2018 | 10:00am | E.2.42

Abstract: We present a comparison between AV1, HEVC and JEM. It reveals that professionally optimized AV1 encoding software is about 32 times as complex as the HEVC encoder software HM. Given the wide attention that AV1 receives and the plans of Google and Facebook to actually use AV1 in commercial environments, the complexity constraints endured during the HEVC development seem to disappear. We present recent results on contour-based prediction for intra coding, a KLT depending on scene and intra prediction direction for coding the prediction error and texture synthesis to replace conventional texture coding for uniform parts of images. Especially contour-based prediction and texture synthesis depend heavily on computer vision algorithms. While typically failures of efficient prediction can be compensated by coding the prediction error and hence a higher data rate, texture synthesis in video coding does not have this safety leash.

Bio: Jörn Ostermann studied Electrical Engineering and Communications Engineering at the University of Hannover and Imperial College London, respectively. He received Dipl.-Ing. and Dr.-Ing. from the University of Hannover in 1988 and 1994, respectively. From 1988 till 1994, he worked as a Research Assistant at the Institut für Theoretische Nachrichtentechnik conducting research in low bit-rate and object-based analysis-synthesis video coding. In 1994 and 1995 he worked in the Visual Communications Research Department at AT&T Bell Labs on video coding. He was a member of Image Processing and Technology Research within AT&T Labs – Research from 1996 to 2003. Since 2003 he is Full Professor and Head of the Institut für Informationsverarbeitung at Leibniz Universität Hannover, Germany. Since 2008, he is the Chair of the Requirements Group of MPEG (ISO/IEC JTC1 SC29 WG11). Jörn was a scholar of the German National Foundation. In 1998, he received the AT&T Standards Recognition Award and the ISO award. He is a Fellow of the IEEE (class of 2005). Joern served as a Distinguished Lecturer of the IEEE CAS Society (2002/2003). He published more than 100 research papers and book chapters. He is coauthor of a graduate level text book on video communications. He holds more than 30 patents.

The review of the TEWI colloquium of Prof. Konstantin Mischaikow from June 6, 2018 comprises the slides (below).

Abstract

With the advent of every improving information technologies, science and engineering is being being evermore guided by data-driven models and large-scale computations.  In this setting, one often is forced to work with models for which the nonlinearities are not derived from first principles and quantitative values for parameters are not known.

With this in mind, I will describe an alternative approach formulated in the language of combinatorics and algebraic topology that is inherently multiscale, amenable to mathematically rigorous results based on discrete descriptions of dynamics, computable, and capable of recovering robust dynamic structures.

To keep the talk grounded, I will discuss the ideas in the context of modeling of gene regulatory networks.

CV

Konstantin Mischaikow earned his Master and PhD degree at the University of Wisconsin–Madison in 1983 and 1985, respectively. Currently he is a Distinguished Professor at Rutgers University, New Jersey, USA. His main research interests are topological methods for the analysis of dynamical systems, computational topology and mathematical biology.  Professor Mischaikow has supervised 16 PhD theses and has been advisor of 21 postdocs. He has over 110 publications, including four books.

He is a leading expert of Conley theory, as well as of rigorous computer-assisted computations. One of his most celebrated results is the proof of chaos in the Lorenz attractor, which serves as a prominent example of the application of both techniques.  In 2014, in recognition of his contributions to dynamical systems as well as to applied and computational topology, Professor Mischaikow was elected to be a fellow of the American Mathematical Society.

The review of the TEWI colloquium of Dr. Hari Prabhat Gupta from June 22, 2018 comprises the slides (below).

Abstract

Internet of Things (IoT) is growing rapidly in decades, various applications came out from academia and industry. IoT is an amazing future to the Internet, but there remain some challenges to IoT for human have never dealt with so many devices and so much amount of data. Machine Learning (ML) is the technique that allows computers to learn from data without being explicitly programmed. Generally, the aim is to make predictions after learning and the process operates by building a model from the given (training) data and then makes predictions based on that model. Machine learning is closely related to artificial intelligence, pattern recognition and computational statistics and has strong relationship with mathematical optimization. In this talk, we focus on ML applications to IoT. Specially, we focus on the existing ML techniques that are suitable for IoT. We also consider the issues and challenges for solving the IoT problems using ML techniques.

CV

Dr. Hari Prabhat Gupta received the B.E. degree in Computer Engineering from Government Engineering College Ajmer, Ajmer, India, the M.Tech. and Ph.D degrees in Computer Science and Engineering from the Indian Institute of Technology Guwahati (IITG), Guwahati, India. He worked with Samsung R&D Bangalore, India. He has received a research fellowship from TATA Consultancy Services, India. He is currently working as Assistant Professor in Department of Computer Science and Engineering, Indian Institute of Technology (BHU), Varanasi, India. His research interests include wireless sensor networks, wireless ad hoc networks, and distributed algorithms. He has published various IEEE and ACM conference papers and Journals in the field of wireless sensor networks.

The review of the TEWI colloquium of Karel Perutka from June 13, 2018 comprises the slides (below) and video here.

Abstract

The main idea of this strategy is based on the essential textbook of pedagogy Orbis sensualium pictus which was written according to Komenský’s belief that the school should be a game. It was first published in Nürnberg in 1658. He believed, contrary to the teaching practices at the time, that the pupils should be able to teach learned matter not only to renounce mechanically but to understand what they were learning. He, therefore, provided a textbook with some illustrations so that it would be captivating for the children. It was about biology (living and inanimate nature), theology and man, something that can now be called the foundations of social sciences.

Modern times have brought new opportunities to implement this strategy.

For example, using simple computer games for practicing and verifying the student’s knowledge. During the lecture, several computer games created for this purpose will be presented. Games are primarily designed to teach automation and programming in MATLAB software at university.

They are created in the way that the data about the matter are read from an external file. This file is enough to be edited and used for any subject of the study program. In the lecture, there will also be introduced several electronic aids facilitating the teaching of work with graphics programs and programs in the office at secondary schools in the Czech Republic. All these games and utilities were rated by students using questionnaires after the completion of the courses, and these results will be presented, too.

Bio

Karel Perutka received his Ph.D. degree in 2007 at Tomas Bata University in Zlin, Faculty of Applied Informatics, Czech Republic, where he is the senior lecturer. Technical Cybernetics was the principal branch of his first research. Karel Perutka was the editor of the book about MATLAB (http://www.intechopen.com/books/matlab-for-engineers-applications-in-control-electrical-engineering-it-and-robotics), the author of one monograph about MATLAB, several book chapters about MATLAB and control theory, and author or co-author of more than 80 papers in the conference proceedings. He is a member of the organizing and reviewing committees of several conferences. He lead more 100 Bachelor and Master Theses.

Karel Perutka is teaching MATLAB programming, electronics, microelectronics, diagnosis of digital systems and modulations and demodulations of signals. He is the most popular teacher of curriculum IT for administrative studiesvoted by students where he teaches the software used in the office.

His main research interests are adaptive control, real-time control, control of multivariable systems, application of MATLAB and new methods of teaching programming and creating didactic aids for secondary schools, programming in C++ and VBA in MS Excel.

He is working on the topic Teaching and practicing the students knowledge using games for last 6 years. He lead the students of Master degree Teachers of Informatics for 6 years, he published 15 papers about this topic, created several teaching games and multimedia tools with the focus in the control theory, the computer graphics, the software in office.

He gave lectures at universities in Europe, mostly in Portugal where he visited ISEP Porto, IST Lisboa, UTAD Vila Real, UA Faro.

He speaks 4 foreign languages. Austria is his favorite country. He goes in Austria also for vacation every year.

Dr. Hari Prabhat Gupta |  June 22, 2018 | 11:00 | Lakeside L4.1.04

Abstract

Internet of Things (IoT) is growing rapidly in decades, various applications came out from academia and industry. IoT is an amazing future to the Internet, but there remain some challenges to IoT for human have never dealt with so many devices and so much amount of data. Machine Learning (ML) is the technique that allows computers to learn from data without being explicitly programmed. Generally, the aim is to make predictions after learning and the process operates by building a model from the given (training) data and then makes predictions based on that model. Machine learning is closely related to artificial intelligence, pattern recognition and computational statistics and has strong relationship with mathematical optimization. In this talk, we focus on ML applications to IoT. Specially, we focus on the existing ML techniques that are suitable for IoT. We also consider the issues and challenges for solving the IoT problems using ML techniques.

CV

Dr. Hari Prabhat Gupta received the B.E. degree in Computer Engineering from Government Engineering College Ajmer, Ajmer, India, the M.Tech. and Ph.D degrees in Computer Science and Engineering from the Indian Institute of Technology Guwahati (IITG), Guwahati, India. He worked with Samsung R&D Bangalore, India. He has received a research fellowship from TATA Consultancy Services, India. He is currently working as Assistant Professor in Department of Computer Science and Engineering, Indian Institute of Technology (BHU), Varanasi, India. His research interests include wireless sensor networks, wireless ad hoc networks, and distributed algorithms. He has published various IEEE and ACM conference papers and Journals in the field of wireless sensor networks.

Omid Esrafilian, MSc |  June 25, 2018 | 3 pm | Lakeside B04, B4.1.114

Abstract

The use of drones, a.k.a. unmanned aerial vehicles (UAVs) as a flying radio access network (RAN) is currently gaining significant attention. It holds promises as a complement to classical fixed infrastructure by allowing ultra flexible deployments, with use cases ranging from disaster recovery scenarios to improving the performance and coverage of the network. Beyond obvious challenges within regulatory, control, navigation, and operational domains, the deployment of autonomous flying-RANs also come with a number of exciting new research problems such as the issue of autonomous real-time placement of the drones in non-trivial propagation scenarios (i.e. scenarios where the optimal placement is not just dictated by a trivial geometry or statistical argument due to shadowing effects, e.g. in cities). We present several different approaches, lying at the cross-roads between machine learning, signal processing, and optimization. Some approaches involve the reconstruction of a city map from sampled radio measurements which can have application beyond the realm of communications.

CV

Omid Esrafilian has started his Ph.D. since 2016 in the Communication System department of EURECOM under the supervision of Professor David Gesbert.

He received both his Master’s and Bachelor’s degree in Control-Electrical Engineering from K.N.Toosi University of Technology in 2016 and 2014, respectively. He graduated as first ranked among M.Sc. students of control major in his university. The subject of His Master’s Thesis was „Simultaneous localization and mapping and autonomous flight of a Quad-Rotor robot using a monocular camera“.

From 2012 to 2016, he worked as a research assistant and head of Quad-Rotor robotic team at the Advanced Robotics and Automated Systems lab (ARAS), K.N.Toosi University of Technology in Iran. During this period, he participated in some robotic competitions and won some prizes.

From 2014 to 2015, he was an assistant professor at Instrumentation Laboratory, K.N.Toosi University of Technology in Iran.

His research interest is robotic (UAVs) and communication and since the beginning of his PhD he published some publications on the topic of UAV-aided Radio Access Networks. Moreover, he was involved in the PERFUM’s first demo of the autonomous cellular flying relays.