VRIP 2019

2019 International Conference on Virtual Reality and Image Processing

Invited Speakers

Prof. Dimitrios Georgakopoulos
Professor and Director of Key Lab for IoT

Swinburne University of Technology, Australia

Biography: Prof. Georgakopoulos is the Director of the IoT Lab at the Digital Innovation Platform of Swinburne University of Technology. Dimitrios came to Swinburne from his roles as Research Director at CSIRO’s ICT Centre and Professor at RMIT. He is also currently the Industry 4.0 Program Leader at Swinburne’s Manufacturing Futures Research Institute and a CSIRO Adjunct Fellow. Dimitrios has held research and management positions in several industrial labs in the USA, including Telcordia Technologies (where he helped found Telcordia’s research laboratories in Austin, Texas and Poznan, Poland), Microelectronics and Computer Corporation, GTE (now Verizon) Laboratories, and Bell Communications Research. Dimitrios is an internationally known expert in IoT, process management, and data management. He has won more than twenty major research awards, produced 200 publications that have been cited 14,400+ times, and attracted significant external research funding ($38M+) from both industry and government in the USA, EU, and Australia.

Title of Speech: From Internet Scale Sensing to Smart IoT-based Solutions  

Abstract: The Internet of Things (IoT) is the latest Internet evolution that incorporates billions of Internet-connected devices that range from sensors, cameras, and wearables, to smart meters, vehicles, and industrial machines. Federations of such IoT devices (which are often deployed and controlled by different providers) can collectively provide high value information and execute appropriate actions that can solve problems that have been too difficult to tackle before. To realize its enormous potential, IoT must provide IoT solutions for discovering needed IoT devices, collecting and integrating their data, analyzing IoT device date and distilling high value information, and performing appropriate actions that impact the physical world, as well as doing all these securely, on the move, and in the cloud or edge. In this talk we provide an overview of the academic and industrial research programs that are being conducted at Swinburne’s IoT lab. We describe the research outcomes that have been developed for a variety of industrial IoT applications in the areas of smart farming, energy, manufacturing, cities, heath and proximity marketing. We also discuss our ongoing fundamental research program in sensor discovery, time-bound IoT data analysis, lightweight IoT security, and Cyber-physical twins that provide the foundation for further generating further research impact. 


Prof. Kok-Leong Ong

La Trobe University, Australia

Biography: Associate Professor Kok-Leong Ong has attracted over $1.5million in research grants related to data analytics research and has served in over 50 Program Committees of international conferences related to analytics, including the prestigious KDD (2015 and 2017) and PAKDD (2006, 2018 and 2019). He is also a Steering Committee Member of AusDM since 2010. Kok-Leong has co-chaired in various roles in different conferences since 2006 and has served as a reviewer for over 20 journals throughout his academic career. He is currently a Section Editor of Business Analytics with the Australasian Journal of Information Systems (ABDC A) and has also served as an industry judge for the Victoria iAwards and the SAP-RMIT University Analytics Competition. Kok-Leong’s research has demonstrated real-life impact. For example, his work on using analytics to improve community well-being was featured by the European Commission in Science for Environmental Policy and many of his papers were ranked by Altmetric to be in the top 25% and 5% of publications. Recently, he was a member of the CRD D2D Education & Training, where he contributed to the design of a national competency framework between 2015 and 2016. He was also one of the chief investigators of a NHMRC Partnership Grant from 2019 – 2023.

Title of Speech: IoT data through social media streams for public health analytics 

Abstract: The public health domain has traditionally operated on survey mechanisms to obtain a snapshot of the public’s tendency towards a number of public health issues, such as maternity health, obesity issues, dietary behaviours, or population health. In this presentation, I will explore how social media and the use of smart phones could be positioned as sources of data streams that can be treated in the same manner as IoT data streams and subsequently used the data for various applications in public health analytics. The presentation will showcase three such applications. The first in population health through community driven noise mapping using smartphones as an IoT device for sound levels. The second application showcases the use of a smartphone app to deliver clinical interventions that captures user behaviours as a data stream of engagement data to deliver personalised and targeted intervention at scale. The last application uses social media data in the context of understanding food trends that is occurring in the public, showcasing how this approach performs better in some applications over traditional census and survey data. 


Assoc. Prof. Simon Fong

University of Macau, Macau SAR

Biography: Simon Fong graduated from La Trobe University, Australia, with a 1st Class Honours BEng. Computer Systems degree and a PhD. Computer Science degree in 1993 and 1998 respectively. Simon is now working as an Associate Professor at the Computer and Information Science Department of the University of Macau. He is a co-founder of the Data Analytics and Collaborative Computing Research Group in the Faculty of Science and Technology. Prior to his academic career, Simon took up various managerial and technical posts, such as systems engineer, IT consultant and e-commerce director in Australia and Asia. Dr. Fong has published over 432 international conference and peer-reviewed journal papers, mostly in the areas of data mining, data stream mining, big data analytics, meta-heuristics optimization algorithms, and their applications. He serves on the editorial boards of the Journal of Network and Computer Applications of Elsevier (I.F. 3.5), IEEE IT Professional Magazine, (I.F. 1.661) and various special issues of SCIE-indexed journals. Simon is also an active researcher with leading positions such as Vice-chair of IEEE Computational Intelligence Society (CIS) Task Force on "Business Intelligence & Knowledge Management", and Vice-director of International Consortium for Optimization and Modelling in Science and Industry (iCOMSI).

Speech Title: White Learning Methodology: A Case Study of Big Medical Data Analysis 

Abstract: Bayesian network is a probabilistic model, which offers a high level of interpretability, non-technical user can intuitively understand the relations of the data. In general, the prediction accuracy may not be one of the highest in the machine learning family. Deep learning on the other hand possess of higher predictive power than many other models. However, one inherent limitation of deep learning is the so-called black-box operations where the representation of prior knowledge is difficult to understand by human users. As a result, many medical practitioners are reductant to use deep learning for critical machine learning application, such as aiding tool for cancer diagnosis. In this paper, a framework of fusing black-box learning that carries the advantage of high level of accuracy, into white-box learning where the prior knowledge and causality information are readily available in a form of direct acyclic graph, is proposed. A case of loosely-coupled white learning model which uses an incremental version of Naïve Bayes network and deep learning is tested on breast cancer diagnosis. White learning is largely defined as a systemic fusion of machine learning models which result in an interpretable Bayes network that explains the relations among the attributes and class; and deep learning model that has superior classification success rate. Our experimental results show that it is possible to create a loosely-couple white-learning model that can do both accurate prediction and data relations reasoning.