Raúl Arrabales Moreno

Cognitive Neuroscience – Artificial Intelligence – Machine Consciousness

Second International Seminar on New Issues in Artificial Intelligence

2nd International Seminar on New Issues in Artificial Intelligence

 February 2 – 6, 2009
Carlos III University of Madrid
Colmenarejo Campus
Avda. Universidad Carlos III, 22
28270 Colmenarejo (Madrid) SPAIN

Organized by ScALAB (CAOS, EVANNAI, GIAA, PLG) Artificial Intelligence Lab at UC3M.

Invited Speakers

Dr. Xin Yao. University of Birmingham.
Evolving Ensemble of Artificial Neural Networks. Co-Evolution, Games and Social Behaviors. Evolutionary Global Optimisation and Constraint Handling.

Dr. Carlos Coello. CINVESTAV-IPN in Mexico.
Recent Results and Open Problems in Evolutionary Multiobjective Optimization.

Dr. Tarunraj Singh. University at Buffalo, The State University of New York.
An overview of advanced estimation algorithms

Dr. Subrata Das. Xerox European Research Centre.
High-Level Information Fusion.

Dr. Michael Buro. University of Alberta, Edmonton, Canada.
Constructing High-Performance AI Systems for Games.

Dr. Malte Helmert. University of Freiburg.
Planning as heuristic search.

Dr. Silvano Cincotti. University of Genoa.
Agent-based Computational Economics.

Dr. Marco DorigoUniversité Libre de Bruxelles.
Bio-inspired Computing: Swarm Intelligence, Ant Colony Optimization and Swarm Robotics.

Speakers and Lectures Details

Dr. Xin Yao, is professor of computer science in the School of Computer Science at the University of Birmingham and the Director of the Centre of Excellence for Research in Computational Intelligence and Applications (CERCIA) which is set up for transfer technology between University of Birmingham and Industry. He is also a Fellow of IEEE, a Distinguished Lecturer of the IEEE Computational Intelligence Society, and a Distinguished Visiting Professor at the Nature Inspired Computation and Applications Laboratory (NICAL) of University of Science and Technology of China, Hefei, China. Dr Yao is the author of numerous journal and conference articles in the area of computational intelligence (evolutionary computation neural network ensembles, computational time complexity of evolutionary algorithm, global optimization and real world applications), and has been implicated in several real applications, working with several industrial partners in Europe. Besides he is the editor or co-editor many books and proceedings, had been invited as plenary speaker at more than 50 conferences, and was the Editor -in-Chief of the IEEE Transactions on Evolutionary Computation from 2003 to 2008 and now become the 2009 Vice President for Publications of the IEEE Computational Intelligence Society.

Institute: University of Birmingham.

Title of lecture: Evolving Ensemble of Artificial Neural Networks. Co-Evolution, Games and Social Behaviors. Evolutionary Global Optimisation and Constraint Handling.

Summary: Designing compact neural networks that generalise well is often very difficult and requires much domain knowledge about the problem. Simulated evolution can be used to search for a near optimal neural network automatically. Since we are going to combine different individual neural networks, the question now is how to design and train those individuals so that they can be combined more effectively. The iterated prisoner’s dilemma (IPD) game has been used extensively in modelling various real-world situations. This talk is concerned with the evolutionary approach to the IPD game, investigating the impact of several aspects as levels of cooperations, reputation. Finally, we present a rigorous theoretical framework of measuring generalisation of co-evolutionary learning quantitatively. Simple evolutionary programming algorithms, called Fast Evolutionary Programming (FEP) and Improved FEP (IFEP), for global optimisation are Analysed to explain why they worked well and when they are unlikely to work. Landscape approximation, cooperative co-evolution and constraint handling techniques — stochastic ranking — are introduced to deal with large and/or complex optimisation problems.

Lecture material: download


Dr. Carlos Coello, received a BSc in Civil Engineering from the Universidad Autonoma de Chiapas in Mexico in 1991. He received a MSc and a PhD in Computer Science in 1993 and 1996, respectively. Dr. Coello has been a Senior Research Fellow in the Plymouth Engineering Design Centre (in England) and a Visiting Professor at DePauw University (in the USA). He is currently full professor at CINVESTAV-IPN in Mexico City, Mexico. He has published over 200 papers in international peer-reviewed journals and conferences. He has also co-authored the book “Evolutionary Algorithms for Solving Multi-Objective Problems” (Second Edition, Springer, 2007) and has co-edited the book “Applications of Multi-Objective Evolutionary Algorithms” (World Scientific, 2004). His publications report over 2,800 citations, half of which are in the ISI Citation Index. Dr. Coello currently serves as associate editor of the “IEEE Transactions on Evolutionary Computation”, “Evolutionary Computation”, “Journal of Heuristics”, “Pattern Analysis and Applications”, “Applied Computational Intelligence and Soft Computing” and “Computational Optimization and Applications”, and as a member of the editorial boards of the journals “Soft Computing”, the “International Journal of Computational Intelligence Research”, “Journal of Memetic Computing”, “Engineering Optimization” and “International Journal of Intelligent Computing and Cybernetics”.

Institute: CINVESTAV-IPN in Mexico.

Title of lecture: Recent Results and Open Problems in Evolutionary Multiobjective Optimization.

Summary: in this lecture, we will review the development of the field known as evolutionary multiobjective optimization (EMO), which is currently a very active research area. This review, which will be provided in a historical way, will be followed by an analysis that aims to answer the following question, which has been posed by several people in the EMO field: will we continue to do research in EMO in the years to come? Several people (e.g., newcomers) have the impression that this field has become more hostile of what it was ten or five years ago, and that it has become much more difficult to make contributions of sufficient value to justify, for example, a PhD thesis. However, a lot of interesting research is still under way. Thus, this lecture will analyze some of the main topics that are currently being investigated in the EMO field, as well as some of the challenges that will be faced in the years to come. This aims to reinforce the hypothesis that this area still has a lot of offer in terms of research and that it has all the elements to remain active for a considerable number of years.

Lecture material: download


Dr. Tarunraj Singh, is a professor in the Department of Mechanical and Aerospace Engineering, State University of New York at Buffalo. He graduated from the Univ. of Waterloo in Canada with a Ph.D in Mechanical Engineering and worked for two years as a post-doctoral fellow at Texas A & M University prior to starting his tenure at SUNY at Buffalo. Dr. Singh teaches undergraduate and graduate courses with a focus on nonlinear control, flexible structure systems and estimation. He has numerous publications in the area of control of slewing flexible structures, target tracking, and data fusion. Dr. Singh is the recipient of the von Humbolt fellowship, SAE Ralph Teetor Educational award and was a NASA Summer Faculty and JSPS fellow. He has been invited to present results of his research at JPL, IBM, Technische Hochshule Darmstadt in Germany and the Aalborg University in Denmark among other places.

Institute: University at Buffalo, The State University of New York.

Title of lecture: An overview of advanced estimation algorithms.

Summary: The seminar is a short overview to advanced data fusion techniques for multisensor multitarget tracking. It will cover first an overview of probabilstic modeling and application of estimation theory for sensor fusion. The classical models (alpha-beta, Kalman filters) will be reviewed, its limitations and introduction to advanced algorithms for nonlinear non-Gausiaan situations (unscented filter, IMM filter, particle filter). The seminar will emphasize the computational aspects and use nonlinear examples where appropriate to illustrate the presented techniques.

Lecture material: download

 

Dr. Subrata Das, is currently reponsible for research at the Xerox European Research Centre in Grenoble, France. Previously he has been the chief scientist at Charles River Analytics, Inc, in Cambridge MA, where he leaded projects in data fusion, decision making under uncertainty, inteligent agents, computational aritifical intelligence and machine learning. Dr. Das is author of numerous journal and conference articles, author/coauthor of three four books and an editorial board memeber of the journal Information Fusion of Elsevier. He received his Ph.D. in computer science from Heriot-Watt university at Scotland, and his M.Tech degree from the Indian Statistical Institute, Kolkata, India.

Institute: Xerox European Research Centre.

Title of lecture: High-Level Information Fusion.

Summary: The seminar is a short introduction to the emerging technology of high-level data fusion, situation assesment and applications. The techniques of information fusion come from different disciplines including uncertainty representation, distributed data processing or artificial intelligence. It will cover a general presentation of information fusion and situation assessment, reviewing the techniques for handling uncertainty and making inferences in distributed situations. It will present some examples of information fusion models and architectures usually employed in application domains such C4I or network centric warfare.

Lecture material: download


Dr. Michael Buro, is an associate professor for computing science at the University of Alberta in Edmonton, Canada. He earned his Ph.D. in 1994 from University of Paderborn in Germany where he studied selective search and machine learning techniques for improving AI systems for two-player games. The result was an Othello playing program that defeated the human World-champion 6-0 in 1997. He is currently conducting research in real-time AI applied to RTS games and sampling based search for imperfect information games. Professor Buro also organizes the annual RTS game AI competition which is based on his free software RTS game engine ORTS.

Institute: University of Alberta, Edmonton, Canada.

Title of lecture : Constructing High-Performance AI Systems for Games.

Summary: Since the inception of AI playing games has been been a popular benchmark. In the past 20 years there have been remarkable advances in the abilities of computer programs to play abstract games such as chess, checkers, backgammon, Othello, and poker at world-championship level. More recently, AI researchers have become interested in developing AI systems for video games as well. In this series of lectures I will cover the following contributions of my research group to the field:
The inner-workings of the Othello program Logistello, which defeated the world-champion 6-0
The development of a card-playing program that recently has reached expert strength in Germany’s national card game skat.
Triangulation-based pathfinding for video games.
The RTSplan planning algorithm for real-time strategy games

 

Lecture material: download


Dr. Malte Helmert, studied Computer Science at the University of Freiburg, Germany, and the University of Durham, England. He earned a Master’s-equivalent degree in 2001 and a doctorate degree in 2006, both at the University of Freiburg. He currently works at that university as a post-doc researcher and lecturer. His main research interests are in the area of classical deterministic planning, focusing in particular on novel planning heuristics and matters of computational complexity. His published work on these topics includes four award-winning papers at major international conferences (ECP 2001, ICAPS 2004, ICAPS 2007 and AAAI 2008).

Institute: University of Freiburg.

Title of lecture: Planning as heuristic search.

Summary: Classical action planning is the problem of finding a sequence of actions that transforms a given initial state into a state that satisfies a given goal. A planning algorithm must be capable of solving all kinds of different search problems without being specifically tailored to any one application domain — a planner must be capable of general problem solving. The scalability of planning systems has increased dramatically in the last decade, with most state-of-the-art algorithms being based on heuristic search. These talks provide an introduction into planning algorithms based on heuristic search, focusing in particular on so-called;relaxation heuristics.

(Remark: the additional talks would then discuss abstraction heuristics.)

Lecture material: download


Dr. Silvano Cincotti, is Head of I2 research group at Department of Biophysical and Electronic Engineering at University of Genoa and also co-founded and heads of the Center for Interdisciplinary Research on Economics and Financial Engineering. His scientific activities have resulted in more 150 scientific articles and more than 50 national and international project grants. He is member of the Society for Computational Economics, of the Society for the Economic Sciences of Heterogeneous Interacting Agents, and of the IEEE within the Computational Finance and Economics Network.

Institute: University of Genoa.

Title of lecture: Agent-based Computational Economics.

Summary: The aim of the seminar is to introduce Agent-based Computational Economics, i.e., the computational study of economic processes modeled as dynamic systems of interacting agents. In the context of ACE, an agent refers broadly to a bundle of data and behavioral methods representing an entity constituting part of a computationally constructed world. Agents can range from active data-gathering decision makers with sophisticated learning capabilities to passive world features with no cognitive function and examples of possible agents include individuals (e.g. consumers, producers), social groupings (e.g. firms, government agencies), institutions (e.g. markets, regulatory systems) and physical entities (e.g. infrastructure, geographical regions). Agent computing represents a methodological approach that could ultimately provide a foundation for economics and other social sciences and the seminar presents the major features of ACE approach with direct reference to economic systems and finance.

Lecture material: download

 

 

 

Dr. Marco Dorigo, is one of the most prominent names in soft computing. He co-directs the Institut de Rechrches Interdisciplinaires et de Développents en Intelligence Artificielle at Université Libre de Bruxelles. He is the author of more than 200 publications and, among his contributions, he invented the Ant Colony Optimization metaheuristic and is at the forefront of the swarm intelligence field.

Institute: Institut de Rechrches Interdisciplinaires et de Développents en Intelligence Artificielle at Université Libre de Bruxelles.

Title of lecture: Bio-inspired Computing: Swarm Intelligence, Ant Colony Optimization and Swarm Robotics.

Summary: The seminar will cover three subjects of Prof. Dorigo’s research agenda. The first one will be a broad introduction to swarm intelligence. Then, he will introduce the Ant Colony Optimization metaheuristic and, finally, he will deal with swarm robotics.

Lecture material: download

Raúl Arrabales

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to top