Professor Luis Martínez University of Jaén, Spain
Group Recommender Systems: A challenge for Consensus Reaching Processes
Due to the fact that in the contemporary E-commerce customers demand quick and easy access to products, and there is an overwhelming amount of information that leads customers into the difficult task of filtering information that meets their actual needs. To address this problem, recommender systems were proposed to filter information, thus delivering to users only the information that meets their preferences or needs. Recommender systems are probably the most successful tool to support personalised recommendations.
Most recommender systems research has been focused on the accuracy improvement of recommendation algorithms. Despite this, recently new trends in recommender systems have become important research topics such as, cold start, context awareness, group recommendations, etc. Group recommendations are very challenging because of its own features, the use of group decision techniques based on consensus reaching processes can provide important improvements an open a line for improvement such a type of recommendations.
This talk will review the basics of the most wide-spread type of Recommender Systems such as Collaborative Filtering models. Afterwards, it will be focused on important results obtained in group recommendations based on group decision negotiation processes. Eventually it will be pointed out different open research lines in the topic.
Luis Martínez was born in 1970. He received the M.Sc. and Ph.D. degrees in Computer Sciences, both from the University of Granada, Spain, in 1993 and 1999, respectively. Currently, he is Full Professor of Computer Science Department at the University of Jaén. His current research interests are fuzzy decision making, linguistic preference modelling, fuzzy systems, decision support systems, personalised marketing, computing with words and recommender systems. He co-edited eleven journal special issues on fuzzy preference modelling, soft computing, linguistic decision making and fuzzy sets theory and has been main researcher in 14 R&D projects, also has published more than 100 papers in journals indexed by the SCI as well as 33 book chapters and more than 150 contributions in International Conferences related to his areas. He is member of the European Society for Fuzzy Logic and Technology, IEEE. Co-Editor in Chief of the of the International Journal of Computational Intelligence Systems and an Associated Editor of the journals IEEE Transactions on Fuzzy Systems, Information Fusion, the International Journal of Fuzzy Systems, Journal of Intelligent & Fuzzy Systems, Applied Artificial Intelligence, Journal of Fuzzy Mathematics and serves as member of the journal Editorial Board of the Journal of Universal Computer Sciences.
He received twice the IEEE Transactions on Fuzzy Systems Outstanding Paper Award 2008 and 2012 (bestowed in 2011 and 2015 respectively). And he is Visiting Professor in University of Technology Sydney, University of Portsmouth Isambard Kingdom Brunel Fellowship Scheme), in the Wuhan University of Technology (Chutian Scholar), Guest Professor in the Southwest Jiaotong University and Honourable professor in Xihua University both in Chengdu (China). Eventually, He received twice the IEEE Transactions on Fuzzy Systems Outstanding Paper Award 2008 and 2012 (bestowed in 2011 and 2015 respectively). And he is Visiting Professor in University of Technology Sydney, University of Portsmouth Isambard Kingdom Brunel Fellowship Scheme), in the Wuhan University of Technology (Chutian Scholar), Guest Professor in the Southwest Jiaotong University and Honourable professor in Xihua University both in Chengdu (China). Eventually, he has been appointed as Highly Cited Researcher 2017 in Computer Science according to the Web of Science.
Dr. Sameer Antani National Library of Medicine, USA
Artificial Intelligence and Clinical Image Analytics for Global Health
For nearly two decades, imaging scientists at the National Library of Medicine (NLM), part of the National Institutes of Health (NIH), have conducted R&D in image and text-based clinical decision support and made significant contributions to state of the art in visual data analytics, machine learning, multimodal information retrieval and data science. This talk will highlight some of the leading projects currently in progress at NLM that apply deep learning and other AI techniques toward these goals. In addition to describing their successes, the talk will also address challenges and considerations in their use for applications in support of clinical decision making. In particular, the talk will focus on aspects of reproducibility and conditions for trust in the AI’s decisions. While recognising the tremendous successes of deep learning and its great potential for human benefit, the talk will consider the impact of data quality, acquisition characteristics, sanity, variety, and volume, and their alignment with the clinical question under consideration.
Dr. Sameer Antani is a computational scientist with the National Library of Medicine, part of the National Institutes of Health. He leads and conducts clinical imaging R&D projects for applications to clinical discovery, data science, and decision support. Examples include use of deep learning techniques for screening for cervical cancer in colposcopic images, and disease staging in histology images, cardiopulmonary diseases in adult and pediatric chest x-rays, detecting malaria in thick and thin red blood smear images, functional brain imaging (fMRI), among others. Dr. Antani maintains an interest in applying artificial intelligence (AI) to medicine particularly for applications in resource-challenged global health settings. He is a Senior Member of the Institute of Electrical and Electronics Engineers (IEEE) and the International Society of Optics and Photonics (SPIE). He serves as Vice Chair on two IEEE committees, the Computer Society’s Technical Committee on Computational Life Sciences (TCCLS), and the Life Sciences Technical Community. Dr. Antani received his doctoral degree in computer science and engineering from the Pennsylvania State University.
Professor Krzysztof Apt Centrum Wiskunde & Informatica, Netherlands
The logic of gossiping
Gossip protocols are concerned with the spread of knowledge in a social network. They aim at arriving, by means of point-to-point or group communications, at a situation in which all agents know each other secrets. These protocols have also been successful in communication networks, in computation of aggregate information, and in a study of data replication.
We consider distributed gossip protocols that are expressed in a simple programming language that employs epistemic logic (for instance in statements such as “if I do not know whether agent i knows my secret I communicate it to him”). These are natural examples of knowledge based programs.
To analyse these protocols, we introduce an appropriate epistemic logic that allows us to discuss the problem of their termination and correctness. We also clarify knowledge-theoretic aspects of communication between pairs of agents, focussing notably on privacy, and indicate their impact on agents knowledge.
The talk is based on joint works with Davide Grossi, Wiebe van der Hoek, and Dominik Wojtczak.
Krzysztof R. Apt is a Fellow at the Center for Mathematics and Computer Science (CWI) in Amsterdam. He holds a PhD degree in mathematics. Apt published four books and more than seventy journal articles, in computer science, mathematical logic and, more recently, economics. He also contributed chapters to more than twenty books and edited several books, all in computer science. His former research dealt with mathematical logic, semantics, verification and design of programming languages, logic and constraint programming, deductive databases and non-monotonic reasoning. His current research is concerned with game theory, social networks, and multi-agent systems. During his scientific career he held tenure positions in Poland,France, USA, and the Netherlands. Apt is a member of Academia Europaea, the founder and first Editor-in-Chief of the ACM Transactions on Computational Logic, past president of the Association for Logic Programming (1997 — 2000), past member of the Executive Committee of the Association for Constraint Programming (2003 — 2006), of the council of the European Association for Theoretical Computer Science (EATCS) (2007 — 2011), and of the Goedel Prize committee (2013 — 2015).
Further, he has been involved since 2001 in a number of open access initiatives, including Computing Research Repository (CoRR) and Electronic Proceedings in Theoretic Computer Science (EPTCS).
Professor Dobrila Petrovic Faculty of Engineering, Environment and Computing, Coventry University, UK
A fuzzy controller for determining return routes in reverse logistics networks
Reverse logistics concerns the return and integration of used or obsolete products into a supply chain. Although the concept of reuse of products and materials is not new, it has been gaining increasing attention in the last decade or so, from various industries, such as automotive manufacturing, electronics and consumer appliance industries. Two main reasons for importance of product recovery are: (1) environmental where new legislation concerning waste reduction requirements, take-back obligations, and disposal practices are introduced, and (2) economical where reverse logistics has been used for an effective reduction of production cost and savings in raw material.
We have considered a generic type of reverse logistics network which consists of a traditional forward production route, inspection of returned products, and three return routes including repair, remanufacturing and disposal routes. The route which a returned product will take depends on its quality. Returned products of better quality can be repaired, and of lower quality can be remanufactured or disposed. We developed a new fuzzy controller which recommends a suitable route which the returned product should take.
Various numerical experiments are carried out to better understand the effects of quality of returns and reverse logistics network parameters on the network performance.
Professor Dobrila Petrovic received BSc and MSc, both in Mathematics/Computer Science, from the University of Belgrade, Yugoslavia, in 1986 and 1991, respectively, and PhD in Engineering, from the University of Warwick, UK, in 1998. She has joined Coventry University, Faculty of Engineering, Environment and Computing, Coventry University, in 1998 and became Professor of Optimisation and Control in 2009.
Her main research areas include modelling and treatment of uncertainty, fuzzy optimisation, fuzzy evolving systems, fuzzy rule-based causal maps in domains such as supply chain management, inventory control, reverse logistics, production scheduling, scheduling in health care and forecasting. Prof Petrovic has carried out research on a number of research projects awarded by the EPSRC (Engineering and Physical Sciences Research Council, UK) in the area of decision support and optimisation of supply chain management and control, FP7 – EU (European Union) grants on supporting highly adaptive network enterprise collaboration and risk management in global production networks, and DSTL (Defence Science and Technology Laboratory) in the area of system modelling under uncertainty.
Prof. Petrovic has published nearly 50 papers in peer reviewed journals and chapters in edited books and nearly 100 papers in proceedings of international conferences. She is an Associate Editor of International Journal of Systems Science, Taylor & Francis, International Journal of Systems Science: Operations & Logistics, Taylor & Francis and IMA Journal of Management Mathematics, Oxford Press.
Dr Hakan Duman Volkswagen Data Lab, Germany
Applying advanced analytics and applied AI to real business questions
With the increasing amount of data now being available, companies in every industry are looking to exploit data for completive advantage, i.e. by understanding customer needs, identifying emerging markets, decreasing production costs and increasing sales etc. Traditional business intelligence solutions are no longer capable of dealing with the huge volume and variety of data produced and captured every day in businesses or made available through ubiquitously interconnected and networked products. However, the rise of computational processing power and the considerable fall of storage technology costs enable the development and utilisation of complex algorithms for profound big data analytics and data mining applications addressing real business questions.
Most common domains for successfully applying big data analytics and applied AI in businesses are in marketing and after sales, e.g. for obtaining a more holistic view of products and customer lifecycles, online advertisement, targeted marketing, as well as in understanding und optimizing business processes and improving business insights. However more opportunities have recently been identified, such as in production, finance, purchasing, logistics etc. where advanced big data analytics bring value to the business and give the organisations the competitive benefit they need.
In the talk, first we will introduce advanced analytics principles then present an excursion on real business questions and discuss how organisations have found many advantages in their explorations with applied AI.
Dr Hakan Duman received his B.Sc. degree in computer science from the Applied University of Regensburg, Germany, and his Ph.D. degree in computer science from the University of Essex, UK, in 1999 and 2008, respectively. He is currently Head of Data Data Science and Applied AI at the Volkswagen Data Lab and Honorary Senior Lecturer at the University of Essex. Prior to that, he worked as Senior Research Scientist at BT Research Labs. He is author of over 30 original research papers in international journals, book chapters, and in international conference proceedings and holds 5 international patents. His research interests include Computational Intelligence, Intelligent Data Analytics / Big Data, Robotics, Ambient and Ubiquitous Intelligence, and Visual Analytics.
Professor Yiannis Demiris Imperial College London, UK
Personal Assistive Robotics
As humans and robots increasingly co-exist in home and rehabilitation settings for extended periods of time, it is crucial to factor in the users’ constantly evolving profiles and adapt the interaction to the personal characteristics of the individuals involved, to move beyond standard shared autonomy paradigms. In this talk, I will describe our computational architectures for enabling human robot interaction in joint tasks, and discuss the related computational problems, including user modelling, attention, perspective taking, prediction of forthcoming states, machine learning, explanation generation, and personalised shared autonomy. I will give some examples from human robot collaboration experiments in robotic wheelchairs for disabled kids and adults, collaboration in musical tasks, activities of daily living (for example dressing tasks), shared control for handheld robots, shared autonomy in driving, among others.
Yiannis Demiris is a Professor of Human-Centred Robotics at the Department of Electrical and Electronic Engineering, Imperial College London, London, U.K, where he heads the Personal Robotics Laboratory. He received the B.Sc. (Hons.) and Ph.D. degrees from the Department of Artificial Intelligence, University of Edinburgh, U.K. His current research interests include human–robot interaction, machine learning, user modelling, and assistive robotics. He has published over 150 journal and peer reviewed conference papers in the above areas. Professor Demiris was a recipient of the Rectors Award for Teaching Excellence in 2012 and the FoE Award for Excellence in Engineering Education in 2012. He is a Fellow of IET, BCS, and the Royal Statistical Society.
Professor Takis Hadjifotiou University College London, UK
High Capacity Optical Fibre Communications Current Status and Future Prospects
During the last twenty-five years optical fibre communications have experienced stupendous growth and they are the foundation of the internet and the World Wide Web. This growth has been based on the simple concepts of direct optical detection and optical amplification. These key featured have enabled the design and manufacture of cost effective systems using wavelength division multiplexing (WDM) with bit rates reaching 100 Gbit/s per wavelength and beyond.
The continuous demand for bandwidth spurted by video on demand, the internet, machine-to-machine communications, IoT, smartphones and the convergence of mobile and fixed networks have forced system designers to consider advanced modulation formats that cannot use direct detection and that entailed the introduction of coherent detection systems.
Coherent detection not only makes possible the use of spectrally efficient modulation formats leading to bit rate in excess of 100 Gbit/s but it does also makes possible the use of digital signal processing, (DSP), and the combination of the two technologies delivers high capacity, spectrally efficient and cost effective optical systems that in terms of information capacity approach the Shannon limit.
This talk will address the question of demand for capacity, the implications for network architectures, the current state of research coherent optical systems and technologies, the use of DSP and the associated technologies, the current state of high capacity systems and it will close with a brief summary of trends for future systems.
After gaining a BSc from Southampton University on communications and technology (1969), a MSc from Manchester University on control systems ( optimum control, stochastic control and aero-space control – 1970) and a PhD from Southampton on digital signal processing for radars (1974), he joined in 1974 Standard Telecommunication Laboratories, (STL), which was the research laboratories of Standard Telephones and Cables, (STC). During his carrier in STL (then from April 1991 in Nortel) he worked on: nonlinear electronic circuits, computer aided design, optical communications (terrestrial, submarine and space), optical technology (sources and detectors), nonlinear system simulation, nonlinear fibre optics. At the end of 2004 he finished his career in Nortel as senior manager (technical director level) responsible for research on future optical systems and technologies.
He has taught postgraduate courses on optical communications and optical technologies in Essex University, University College London, University of Bangor, University of Heriot Watt and St Andrews, University College Cork (Republic of Ireland), Dublin City University (Republic of Ireland) and University of the Peloponnese (Hellas). He is currently Visiting Professor of Bangor University and University College London. He is a Fellow of the Royal Academy of Engineering (RAEng) and of the Institution of Engineering Technology (IET), and member of the Institute of Physics (IoP), IEEE, and the Optical Society of America.
Professor Reinhold Scherer University Institute of Neural Engineering at the Graz University of Technology, Austria
Brain-Computer Interfaces for end-users with cerebral palsy: Lessons learned
Cerebral palsy (CP) is a non-progressive condition caused by damage to the brain during early developmental stages. Individuals with CP may have a range of problems related to motor control, speech, comprehension, or mental retardation. Many children with CP have normal intelligence, however, due to lack of appropriate communication means they are classified as educational abnormal. Communication solutions are available; however, they strongly depend on motor activity and assistance of others.
The aim of the EC-FP-7 ABC project was to develop an interface for individuals with CP that improves independent interaction, enhances non-verbal communication and allows expressing and managing emotions. Both motor activity and (electro)physiological signals will serve as input for the interface. In fact, one key component of the ABC system is a non-invasive electroencephalogram-based (EEG) brain-computer interface (BCI). This talk gives an overview of the implemented BCI system and outlines issues encountered during development. Challenges include user education, training paradigms, signal processing and realistic assessment of performance.
Reinhold Scherer is Associate Professor and Deputy Head of the Institute of Neural Engineering at the Graz University of Technology, Austria. He is member of the Laboratory for Brain-Computer Interfaces (BCI-Lab) at Graz University of Technology and of the Institute for Neurological Rehabilitation and Research at the rehabilitation center Judendorf-Strassengel, Austria.
In 2008 he received his PhD in computer science from Graz University of Technology, where, beginning in 2001, he worked on non-invasive electroencephalogram-based (EEG) brain-computer interfacing (BCI). He spend the years from 2008 to 2010 as postdoctoral researcher at the Department for Computer Science & Engineering, University of Washington, Seattle, USA, and was member of the Neural Systems and the Neurobotics Laboratories at the University of Washington.