Prof. Marko Čepin

University of Ljubljana, Faculty of Electrical Engineering, Ljubljana, Slovenia

Reliability of Smart Grids

A smart grid is a term representing an electrical power system, which efficiently deliver sustainable, economic and secure supply of electric energy of required quality to its customers employing innovative products and services together with intelligent monitoring, control, communication, and self-healing technologies. It is realised with the increased use of digital information and controls technology to improve reliability, security, and efficiency of the electric power system including dynamic optimization of system operations and distributed resources, with full cyber-security. The term smart grid includes the real-time, automated and interactive technologies that optimize the physical operation of appliances and consumer devices for metering and communications concerning the power system operation and status. It includes the advanced electrical energy storage and peak-shaving technologies including plug-in electric and hybrid electric vehicles. The system is far more complex in many layers compared to the traditional power system and thus brings more and more challenges to its reliable operation. The issues related with evaluation of the power system reliability are reviewed in sense of their applicability for evaluation of future smart grids. Traditional and advanced reliability measures are discussed The focus is placed to those related with the intermittent power generation sources which are more and more distributed. The discussion concludes with some findings including the fact that more reliable systems and subsystems face smaller probabilities of faults, but even rare faults can cause much larger undesired consequences than before because the systems are more and more complex.

Prof. Coen van Gulijk

Institute of Railway Research, University of Huddersfield, United Kingdom

Digital re-engineering for safety management

In a time of accelerated introduction of digital systems, the demand for better safety increases as well. To keep up with modern developments digital re-engineering of Safety Management Systems is inevitable. However, the development of interactive online safety decision support systems is not straightforward and safety staff are not well trained in digital techniques. This presentation will shed light on the challenges of digital re-engineering of safety management systems and the tasks that lie ahead for safety researchers.

Prof. Min Xie

Chair Professor, City University of Hong Kong, Hong Kong

Opportunities and Challenges of the Reliability of AI Systems

Artificial intelligence (AI) is being incorporated into many complex systems and it is a popular topic for research and application today. Many people talk about different AI methodologies and application examples in different contexts. However, there are numerous challenging problems in different AI applications. How to measure the reliability of AI is unclear to begin with, not to say how to develop reliable AI technology and systems. In this talk, we will share some thoughts and experience from systems engineering perspective and discuss some of these issues. Some of the research related to software reliability, hardware reliability and combined systems can be strongly linked to this. AI applications also rely on the availability of large amount of data that may have other issues such as data uncertainty and integration.

Prof. Soda Paolo

Campus Bio-Medico University in Roma, Italy

Artificial intelligence meets clinical imaging for personalized medicine in oncology

High throughput standard-of-care medical images such as CT, PET or MRI are now used in oncology to detect lesions, to plan treatments, to follow up the disease, etc. Furthermore, the increasing adoption of electronic patient records as well as the diffused use of PACS have made available heterogeneous patient data, spanning different spatial and temporal scales, modalities, and functionalities. The quantitative nature of the images allows us to go beyond visual interpretation by computing, analysing and selecting advanced quantitative imaging features. This in turns has led to radiomics, which is also evolving into radiogenomics since it is looking for correlation between cancer imaging features and gene expression. All this quantitative information can be harnessed in an integrated platform and leveraged via clinical-decision support systems (CDSSs) by artificial intelligence to improve personalized medical decision-making with diagnostic, prognostic or predictive value.. How to measure the reliability of AI is unclear to begin with, not to say how to develop reliable AI technology and systems. In this talk, we will share some thoughts and experience from systems engineering perspective and discuss some of these issues. Some of the research related to software reliability, hardware reliability and combined systems can be strongly linked to this. AI applications also rely on the availability of large amount of data that may have other issues such as data uncertainty and integration.