Keynote speakers from: 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 | 2011
Dr. Veselin Raychev
View Google Scholar Profile
Keynote on: Beyond Neural Networks for Programming: How to Learn from Programs
Biography: Dr. Veselin Raychev obtained his PhD from ETH Zurich in 2016 in the area of machine learning for programs. His doctoral dissertation on this topic was honored by the ACM as one of the best three dissertations in computer science worldwide in 2016. Prior to his PhD, Veselin worked for five years as a software engineer at Google.
Prof. Valeri Mladenov
Technical University of Sofia, Bulgaria
Keynote on: Control of various robots through signals from the brain activity
Biography: Valeri Mladenov received his Ph.D. from Technical University of Sofia (TU Sofia), Bulgaria in 1993. In 2019 he defense a "Doctor of Sciences" thesis at the same institution. In 2004 he becomes a Head of the Department of Theory of Electrical Engineering. Since June 2011 he was a Dean of Faculty of Automation, since Dec. 2011 he has been a Vice-Rector of TU Sofia and since Dec. 2015 he is a Director of the directorate of Information and public relations. He is a guest lecturer at the Faculty of Electrical Engineering, Eindhoven University of Technology, in the Netherlands and many others.
In 2014 he has been a Deputy Minister of Education and Science in the Caretaker Government in Bulgaria.
Prof. Mladenov's research interests are in the field of nonlinear circuits and systems, neural networks, artificial intelligence, applied mathematics and signal processing. He has received many international research fellowships. He has more than 300 scientific papers in professional journals and conferences. He is a co-author of ten books and manuals for students. He had received many research grants from the Technical University of Sofia, Bulgarian Ministry of Education and Science, DAAD - Germany, NWO - Netherlands, Royal Society - UK, NATO, TEMPUS, Erasmus and others and also with his team he participated and participate in many national and international projects - H2020, FP7, DFG, Erasmus+ and others.
As a member of several editorial boards, Prof. Mladenov serves as an editor in chief, associate editor, and reviewer for a number of professional journals and conferences. He is a Senior Member of IEEE, a member of the IEEE Circuit and Systems Technical Committee on Cellular Nanoscale Networks and Array Computing and Educational Activities Officer of the Bulgarian IEEE section. He is also a member of the International Neural Networks Society (INNS), member of the International Council of Large Electric Systems, (SIGRE), member of the Steering Committee of the International Symposium on Theoretical Electrical Engineering (ISTET), a member of the Management Boards of the Scientific and Technical Union of the Power Engineers, and the Union of Automation and Informatics in Bulgaria.
Assoc. Prof. Dr. Plamen Vatchkov
National Center for Supercomputer Applications, Sofia, Bulgaria (NCSA)
Keynote on: European and Bulgarian Supercomputers
Biography: Plamen Vatchkov received his Ph.D. from Moscow Power Institute in 1977.In 1983 he became Deputy Director of the Institute of Technical Cybernetics and Robotics, Bulgarian Academy of Sciences. In 1986 he received the degree Senior research fellow (Associate Professor) in the same institute. During the period 1984_1992 he was General Director of the company “Microprocessor Systems”-Pravetz. Between 1992 and 2005 Plamen Vatchkov was Manager in different Bulgarian and international IT end Telecom companies: Multitech AD, GloboAD, Cable Bulgaria AD, Overgaz AD end Cabletel AD. During the period 2005-2009 he was at the position Chairman of the State Agency of Information Technologies end Telecommunications (Minister of Information Technologies and Telecommunications). Between 2010-2012 he was the General Secretary of the Federation of the Scientific and Technical Unions (FNTS). From 2008 untill now Plamen Vatchkov is Vice Chairman of NCSA and Sofia region organization of FNTS.
Assoc.Prof. Vachkov has research interests in the field of IT, PC, Supercomputers and power electronics. He has several specializations in Microprocessor devises (France), Management (USA) and Quality management (Japan). He has published more than 40 articles and one book in the field of IT and electronics. He was Chairman of the Council of International Telecommunications Union of UN (2009). Nowadays he is Chairman of the organizing committee of Annual Bulgarian IEEE conference Telecom.
Beyond Neural Networks for Programming: How to Learn from Programs
As the size of open source code has grown dramatically, so has the number of machine learning tools that propose to solve software engineering tasks by learning from these codebases. Yet, it has turned out that producing tools which actually work in the real world is harder than expected with even the latest neural network techniques failing to provide reasonable solutions.
We propose a method to address this problem via two main ideas: (i) we leverage static analysis techniques to extract semantic representations of programs and build our models over the program semantics and not directly over their syntax, (ii) importantly, we ensure the results of the probabilistic model are limited to explainable and semantically interesting predictions.
We will illustrate these concepts in action by demonstrating several production-quality software systems (e.g., state-of-the-art taint-based security bug-finding tool) running at DeepCode.
Control of various robots through signals from the brain activity
Prof. Valeri Mladenov
The research of the brain and how it works is a hot topic and gaining popularity in recent years. Therefore, it is not surprising that more and more researchers, as well as young scientists, are interested in trying to connect different hardware with brain activity – mind-controlled cars, planes, robots. In this presentation, a brief overview of how to record and proceed with the brain activity signals will be given. Special attention will be done on the developed brain activity management system that can successfully control “Baxter” and “Nao v5” robots. The system uses the Emotiv Epoc+ electroencephalograph (EEG) to capture real-time brain activity. The hardware follows the International 10-20 Electrode Positioning System.
In order to achieve control with the mind, a user must create their own profile. This profile stores that user’s specific personal brain activity data because each person has very unique brain activity. First, a neutral level of brain activity must be recorded and then it is necessary to train the various commands to the robots. The current version of the system allows distinguishing between 5 different commands. The first command is relatively easy to train. But with each other successfully trained command it becomes more difficult to distinguish the individual command from the others. Training is done through a virtual cube, so over a period of time, a person can learn to control their brain activity. It can be moved in different directions, and if a person has a highly developed imagination, they can even make the cube disappear.
The EEG data from the device is processed and compared with that of the user’s profile. When there is a coincidence between both the data from the device and the data already trained by the user, the command is sent to the robot in real-time for execution. Specially written Python scripts are used to make the connection between Emotive Epoc + and the robots.
Each robot can perform up to 5 different commands individually - four-movement commands and one command to return the robots to a neutral position. The results depend mainly on how well the system is trained and how long the training is done. With enough training time and the appropriate training methods, each of the five commands can be distinguished and executed. For comparison on average, most users manage to train up to 3 different commands correctly.
European and Bulgarian Supercomputers
Assoc. Prof. Dr. Plamen Vatchkov
Supercomputers, known as well as high performing computers (HPC) recently become a very important part of computer technology, contemporary industries and science. The competition between the producers, the scientific researchers and countries in this field became a key factor of the human progress. European commission lunch an initiative PRACE (Partnership for Advanced Computing in Europe), as an instrument for cooperation of European countries. Today 47 Universities and Research centers participate in this initiative. The article focuses on the existing European supercomputers and the main characteristics of the future project. The history of supercomputers in Bulgaria is presented. The examples of the new project for petascale computers in the host countries are described. Some key examples for different applications of supercomputers in naval and automotive industries and pharmacy are given.