Ethical Design of Large-Scale Agent-Based Systems
Seminar title: Ethical Design of Large-Scale Agent-Based Systems
Speaker: Professor Yves Demazeau, Centre National de la Recherche Scientifique, Laboratoire d’Informatique de Grenoble, France.
Seminar room: CB11.12.113
Date and time: 12:00 – 13:00, March 31, 2016
Date and time: 12:00 – 13:00, March 31, 2016
Seminar chairman: Dr Guandong Xu
Abstract: In order of increased focus on end users, engineering multi-agent systems must evolve towards more ethical systems, in the respect of the end users these systems are designed to serve. We propose an iterative method of constructing such ethical systems, which is divided into six stages, independently verified and validated as a whole.
We then focus on the two first stages of the method: 1. Modelling end-users in terms of preferences, capabilities, activities and expectations, and the environment in which they operate. 2. Protecting personal end user data by mechanisms to ensure privacy.
We investigate how to determine, at design time, the right balance between legal aspects and moral aspects of such agents. We show how this balance depends in general on the problem to solve, of the application domain, and how it should evolve over time.
We illustrate the design of such large-scale agent-based systems through four application domains: office assistants, autonomous transports, health informatics. We discuss how much the balancing differs from one country to another and how things might evolve in the future.
We believe the methodology of designing large-scale agent-based systems we propose can be agreed internationally to enable the sharing of the systems at a larger extend. With this work, we believe to pave the way to the development of more sophisticated autonomous agents and multi-agent systems that are “ethical-by-design”, for the benefit of end users.
Biography of speaker:
Yves Demazeau received his PhD in Computer Science at Grenoble National Polytechnic Institute in 1986. He is Director of Research at CNRS, and External Professor at University of Southern Denmark since 2002. He has been Visiting Professor at VUB Brussels in 1989, Southern Denmark University in 1994, and University of Technology, Sydney in 2014. He has edited 28 books or proceedings, has authored or co-authored 173 papers and has given 46 invited lectures, 143 seminars, and 22 tutorials. He has advised 51 MSc and 27 PhD theses.
Yves Demazeau is one of the co-founders of the Multi-Agent Systems (MAS), one of the 10 main subdomains of Artificial Intelligence. He has been active in this area for 28 years. He is also well known for the large-scale and original applications he has developed in various domains: autonomous robotics, bio-inspired robotics, computer vision systems, energy management, geographical information, interactive games, natural language processing, personal assistance, support for artistic creation, telecommunications, and web Intelligence.
During these last ten years, Yves Demazeau has expanded his research to more general Artificial Intelligence issues. His current basic research activities in Artificial Intelligence address four themes: end-user centration, from individuals to groups; privacy issues and trust concerns; observation and evaluation of large-scale systems; ethical design of large-scale agent-based systems. His current favourite application domains are also at the number of four: service to the person, bio-inspired robotics, energy management, and financial modelling.
Yves Demazeau has been teaching MAS at academic level in Western Europe (AGH Krakow, UJF Grenoble, UL Luxembourg, UPD Paris, UPM Madrid, USD Odense), and in Southern America (UFC Fortaleza, UFRGS Porto Alegre, UNICEN Tandil) as well as in International Summer Schools such as IBERAMIA and IJCAI. Creator of MAAMAW, CEEMAS, and ICMAS conferences, he is now orchestrating the PAAMS series since 2009 and is supporting the BESC series since 2015. He has participated in 253 conference program committee.
Yves Demazeau has served or is serving as editorial board member of several Journals (JAAMAS Springer Verlag, MAGS IOS Press) and he is the Editor-in-Chief of the French AI Journal (RIA Hermes) since 2010. He is President of the French Association for Artificial Intelligence (AFIA) since 2011 and he is orchestrating the AI events in France like RFIA to be held in Clermont-Ferrand in the summer of 2016. He has held consulting activities in the past with Thomson-CSF , FRAMENTEC-Cognitech, Unilog and currently with A2I Systems.