• The 11th International ACM Conference on Management of Digital EcoSystems (MEDES'19) - 12-14 November 2019, Limassol - Cyprus

  • Paper Submission Deadline: April 30, 2019

  • Author Notification: June 14, 2019

  • Venue: Miramare Hotel

Visual Analytics Framework:
For enhanced decision making using cascaded AI technologies
Talk 2 Title
key1
Dimitrios Tzovaras

mar15

Marios Dikaiakos

  

A Visual Analytics Framework for enhanced decision making using cascaded AI technologies

key1

 by Dimitrios Tzovaras

Center for Research and Technology Hellas - Information Technologies Institute
Greece

Abstract

Taking advantage of the massive amounts of processing power offered by modern CPUs & GPUs, state-of-the-art AI is rapidly moving from traditional pattern recognition and machine learning technologies to deep neural architectures.

Despite the impressive performance of the latter, they are still able to carry out only very specific operations, with limited configuration margin, esp. when their training has been completed.

Their, yet undiscovered to its full extent, potential for unprecedented accuracy lies within the depth of their structure, i.e. the large number of layers, that achieves to encode paramount intelligence tailored to the exact needs of their objective; always upon correct and adequate training.

Inspired by the effectiveness of such multi-layered approaches, the current session will introduce a framework implemented as a cascaded architecture of Deep Neural Networks (DNNs) that aims to implement Decision Making Support with efficient applicability in a wide range of fields of applications. Task-specific DNNs lie in the basis of the pyramid, while the DNNs found higher in the hierarchy act as "control gates" for the ones below, while seamlessly offering "dimensionality reduction", "data minimization" & "data mining" through fusion, classification & clustering techniques.

Yet, of both extreme operational & scientific interest is the involvement of the human operator in the loop, so as to benefit from their expertise and to interactively serve their preferences & needs per case. In order to achieve this, the crown of the proposed framework consists of a Visual Analytics toolset that not only allows for comprehensive visualizations of the produced results/outcomes, but it also offers advanced interaction possibilities. This way, a direct link to the aforementioned control-gates is established, transmitting the appropriate operational configuration to the lower processing layers, forming thus, the desired communication loop between human intelligence & AI.

The applicability of the proposed framework is demonstrated in a series of real-world use cases that involve a series of distinct challenging tasks but also the intensive feedback by each operator.

In particular, they range from (i) the detection & mitigation of cybersecurity attacks in a real-world IoT networks and (ii) early event detection from social media & news feeds, to (iii) real-time defect detection on materials in the production line and (iv) DNA sequences analysis.

Biography

Dr. Dimitrios Tzovaras is a Senior Researcher Grade A’ (Professor) and the Director at CERTH/ITI (the Information Technologies Institute of the Centre for Research and Technology Hellas). He received the Diploma in Electrical Engineering and the Ph.D. in 2D and 3D Image Compression from the Aristotle University of Thessaloniki, Greece in 1992 and 1997, respectively. Prior to his current position, he was a Senior Researcher on the Information Processing Laboratory at the Electrical and Computer Engineering Department of the Aristotle University of Thessaloniki. His main research interests include visual analytics, computer vision, data fusion, machine learning and artificial intelligence, network and computer security, biometrics and virtual reality. He is author or co-author of over 130 articles in refereed journals and over 350 papers in international conferences.

Since 2004, he has been Associate Editor in the following International journals: Journal of Applied Signal Processing (JASP) and Journal on Advances in Multimedia of EURASΙP. Additionally, he was an Associate Editor in the IEEE Signal Processing Letters journal (2009-2012) and Senior Associate Editor in the IEEE Signal Processing Letters journal (2012-2014), while since mid 2012 he has been also Associate Editor in the IEEE Transactions on Image Processing journal. He is currently a Senior Associate Editor in the IEEE Transactions on Image Processing journal.  Over the same period, Dr. Tzovaras acted as ad hoc reviewer for a large number of International Journals and Magazines such as IEEE, ACM, Elsevier and EURASIP, as well as International Scientific Conferences (ICIP, EUSIPCO, CVPR, etc.).

Since 1992, Dr. Tzovaras has been involved in more than 100 European projects, funded by the EC and the Greek Ministry of Research and Technology. Within these research projects, he has acted as the Scientific Responsible of the research group of CERTH/ITI, but also as the Coordinator and/or the Technical/Scientific Manager of many of them (coordinator of technical manager in 24 projects – 12 H2020, 1 FP7 ICT IP, 7 FP7 ICT STREP, 3 FP6 IST STREP and 1 Nationally funded project).

 

 

 

 

Talk 2 Title

mar15

by Marios Dikaiakos

Departement of Computer Science - University of Cyprus
Cyprus

Abstract

Abstract goes here

Biography

Marios D. Dikaiakos is Professor of Computer Science at the University of Cyprus. He serves as Director of the Center for Entrepreneurship of the University, since January 2015. He was Head of the Computer Science Department between 2010-2014. He is also the Founding Director of the Laboratory for Internet Computing. Dikaiakos received his Ph.D. in Computer Science from Princeton University (1994), an M.A. degree from Princeton (1991), and a Dipl.-Ing. degree from the National Technical University of Athens (summa cum laude, 1988). He was a Research Associate at the University of Washington in Seattle (1994-1995) and has held short-term visiting positions at the Paris Research Lab of Digital Equipment Corporation, the University of Crete, and Rutgers University.

Dikaiakos' research focuses on Internet Computing. Recent research activities focus on Cloud Computing, Online Social Networks, and Vehicular Computing. He has been a principal institutional investigator or principal investigator for 22 projects funded by the European Union and the Research Promotion Foundation of Cyprus, and has published over 170 papers in books, international scientific journals and refereed conference proceedings, and was in charge of the development of several research software systems released internationally. More information on ongoing projects, open research positions, past and future events can be found on the Web site of the Laboratory for Internet Computing.

Interesting Links