Home      Log In      Contacts      FAQs      INSTICC Portal
 

Keynote Lectures

The Two Pillars
Robert Pergl, Czech Technical University in Prague, Czech Republic

Digitalization: A Meeting Point of Knowledge Management and Enterprise Engineering
Eduard Babkin, National Research University “Higher School of Economics”, Russian Federation

To Combine or Not to Combine: Ranking and Scoring for Data Analytics and Knowledge Discovery
Frank Hsu, Fordham University, United States

Kandinsky Patterns
Heimo Müller, Medical University of Graz, Austria

 

The Two Pillars

Robert Pergl
Czech Technical University in Prague
Czech Republic
 

Brief Bio
Dr. Robert Pergl is an Associate Professor at Department of Software Engineering, Faculty of Information Technologies of Czech Technical University in Prague, Czech Republic, where he founded "Centre for Conceptual Modelling and Implementation", a group focusing on research, development and applications of methods and tools for ontological engineering, enterprise engineering, software engineering and data stewardship. Apart from his publishing work, Dr. Pergl is also a General Chair of EOMAS Workshop, a representative in the CIAO! Enterprise Engineering Network and National Node Committee member of ELIXIR Czech Republic.


Abstract
In his keynote, Dr. Pergl is going to discuss two pillars of intellectual human endeavour: naming and hierarchies. He digs into the essence of these corner-stones of conceptualisation and explores their presence, significance and forms in various disciplines. Challenges of naming and hierarchies in engineering disciplines are discussed and "lessons learned" are formulated.



 

 

Digitalization: A Meeting Point of Knowledge Management and Enterprise Engineering

Eduard Babkin
National Research University “Higher School of Economics”
Russian Federation
 

Brief Bio
Eduard Babkin is a tenured professor in the department of Information Systems and Technologies of National Research University Higher School of Economics (HSE Nizhny Novgorod, Russia), where he takes a position of the head in the research laboratory of theory and practice of decision support systems (TAPRADESS). Also Eduard Babkin has been working in IT industry, he has more than twenty years of practical experience in architecting, software design and project management of complex distributed information systems. In 1993 Eduard Babkin obtained BS degree in Informatics and started the academic carrier as a lecturer. In 2007 Eduard Babkin obtained his PhD degree in Computer Science in National Institute of Applied Sciences (Rouen, France). Since that time he has been carried out scientific research in enterprise engineering, multi-agent systems, knowledge management as a principal investigator or a team lead. Currently Eduard Babkin is mostly interested in multidisciplinary studies where advances of conceptual modeling, distributed algorithms and multi-agent systems were fused with corresponding domains of sociology and economics.


Abstract
Digital transformation of organizations became a significant research and engineering challenge worldwide. In many cases digitalization requires extraction of tacit individual, interpersonal or organizational knowledge to explicit machine-readable forms and their conscious application during enterprise reengineering. Successful accomplishment of these tasks vitally relies on a rigorous scientific theory and formal methods. This lecture demonstrates how the technique of evolvable domain-specific languages solves several problems of knowledge management in organizations, the enterprise ontology approach facilitates comprehensive understanding of socio-technical systems, and how fusion of these approaches may provide a reliable tool for digitalization. Illustration of results obtained in several research projects supports the main statements of the lecture.



 

 

To Combine or Not to Combine: Ranking and Scoring for Data Analytics and Knowledge Discovery

Frank Hsu
Fordham University
United States
 

Brief Bio
D. Frank Hsu is the Clavius Distinguished Professor of Science, a professor of Computer and Information Science, and director of the Laboratory of Informatics and Data Mining at Fordham University in New York, USA. He was chair of the CIS Department and associate dean of the Graduate School of Arts and Sciences. He held visiting positions at JAIST, Keio University, MIT, Taiwan University, and University of Paris-Sud. Hsu’s main research interests are: interconnection networks, graph database, micro- and macro-informatics, data science, ensemble method, and combinatorial fusion algorithm. He has co-authored/co-edited 40 books and book chapters and published over 200 technical papers. He has given over 400 presentations worldwide. Hsu served or is serving on many editorial boards including IEEE Transactions on Computers, IEEE Transactions on Reliability, IEEE Systems Journal, Brain Informatics, and Journal of Interconnection Networks. Among the honors and awards he received are IEEE-AINA Conference Best Paper Award, Foundation Fellow of ICA, Fellow of ICIC, Fellow of the New York Academy of Sciences; and IBM Faculty Award. Hsu received his M.S. from the University of Texas at El Paso and Ph.D. from the University of Michigan. He is a Senior member of the IEEE. (http://storm.cis.fordham.edu/~hsu)


Abstract
In the big data era, scientific discovery of knowledge tends to have fewer but sophisticated experiments with more variables (cues, criteria, features, attributes, or indicators) and larger number of hypotheses. As such, various ensemble methods combining multiple models or multiple machine learning algorithms are frequently used to improve forecasting, prediction, decision making, and policy formulation. However, it remains to be a great challenge to know when and how to combine these models or algorithms. This keynote talk will cover the design of intelligent scoring systems and discuss when and how these systems should be combined using a rank-score characteristic (RSC) function and the notion of cognitive diversity. Examples will include figure skating judgement, information retrieval systems, intrusion detection, wireless network selection, and multi-layer combinational fusion.



 

 

Kandinsky Patterns

Heimo Müller
Medical University of Graz
Austria
 

Brief Bio
Heimo Müller, born in Austria not far from the Slovenian Carinthian border, studied mathematics in Graz and Vienna. He began his professional career in computer graphics and multimedia at Joanneum Research at the Institute for Digital Image Processing and Computer Graphics and at the Institute for Information Systems. His work in the field of film and video, including storage, indexing, archiving and restoration, is particularly noteworthy. As Marie-Curie Research Fellow at the Free University of Amsterdam he was involved in the modelling of semantic structures in moving image sequences and back in Graz Heimo Müller was the founding head of the Information Design course at the University of Applied Sciences Joanneum. Since a decade he is at the Medical University of Graz working on data management in biobanka and precision medicine. In Particular his research topics today are  visual computing, information design, digital pathology and – most important – explanability of AI in the medical domain. 


Abstract
AI follows the notion of human intelligence, which is unfortunately not a clearly defined term. The most common definition, as given by cognitive  science as mental  capability,  includes,  among  others,  the ability to think abstract, to reason, and to solve problems from the real world. A hot topic in current AI/machine learning research is to find out whether and to what extent algorithms are able to learn abstract thinking and reasoning similarly as humans can do – or whether the learning out-come remains on purely statistical correlation. We introduce so-called Kandinsky Patterns as a framework to compare human and machine "thinking". Kandinsky Patterns are mathematically describable, simple,  self-contained  hence controllable test data sets for the development, validation and training of explainability in AI. Kandinsky Patterns are at the same time easily distinguishable from human observers. Consequently, controlled patterns can be described by both humans and computers.



footer