• Russell C. Eberhart

• Xin Yao

• Hiroshi Mamitsuka

• David G. Stork

• Bailin Hao
  

Swarm Intelligence and Evolutionary Analog Computing

Russell C. Eberhart, Professor & Ph.D, Fellows of the IEEE and the AIMBE

Electrical and Computer Engineering at the Purdue School of Engineering and Technology, Indiana University Purdue University Indianapolis (IUPUI), Indiana, USA
Email: reberhar@iupui.edu
Personal website: http://www.engr.iupui.edu/~eberhart/

Abstract: First, recent developments in swarm intelligence and particle swarm optimization are reviewed. The evolution of neural networks using swarm intelligence is discussed. Next, a brief overview of the relatively new field of extended analog computing is presented, highlighting the differences between traditional analog computing and extended analog computing. The use of swarm intelligence to evolve configurations of extended analog computers, resulting in evolutionary analog computers (EvACs) is described. A new EAC supercomputer project is described. At the conclusion of the plenary presentation, the evolution of an EvAC will be demonstrated via the Internet using swarm intelligence and an extended analog computer.

Bio-Sketch: Russell C. Eberhart (http://www.engr.iupui.edu/~eberhart/) is Professor of Electrical and Computer Engineering at the Purdue School of Engineering and Technology, Indiana University Purdue University Indianapolis (IUPUI). He is also Vice President and Chief Technology Officer of Computelligence LLC, Indianapolis, Indiana. He received his Ph.D. from Kansas State University in electrical engineering. He is co-editor of a book on neural networks, and co-author of Computational Intelligence PC Tools, published in 1996 by Academic Press. He is co-author of a book with Jim Kennedy and Yuhui Shi entitled Swarm Intelligence, published by Morgan Kaufmann/Academic Press in April 2001. He was awarded the IEEE Third Millenium Medal. In 2001, he became a Fellow of the IEEE, and in 2002 he became a Fellow of the American Institute for Medical and Biological Engineering. He is the co-author, with Yuhui Shi, of a book entitled Computational Intelligence: Concepts to Implementations, published by Morgan Kaufmann/Elsevier in 2007. His areas of research include swarm intelligence and extended analog computing, and the detection of sleepy and inattentive driving.

Why Evolutionary Computation

Xin Yao, Professor & Ph D, Fellow of the IEEE, EIC of IEEE TEC
Department of Computer Science, the University of Birmingham, UK
Email: X.Yao@cs.bham.ac.uk
Personal website: http://www.cs.bham.ac.uk/~xin/

Abstract: Evolutionary computation is the study of computational systems that get inspirations and ideas from natural evolution and other biological systems. It has been widely used in optimization, learning and creative design with success. There has been a steady improvement of our understanding of the theoretical foundations of evolutionary computation in recent years. This talk will review some of the recent developments in evolutionary optimization, evolutionary learning and computational time complexity of evolutionary algorithms. It is pointed out that evolutionary computation is an extremely rich research field, and certainly much more than any single algorithms.

Bio-Sketch: Xin Yao (http://www.cs.bham.ac.uk/~xin/) is a Professor (Chair) of Computer Science at the University of Birmingham, UK. He obtained his BSc from the University of Science and Technology of China (USTC) in Hefei, China, in 1982, MSc from the North China Institute of Computing Technology in Beijing, China, in 1985, and PhD from USTC in Hefei, China, in 1990.
He was a postdoctoral research fellow at the Australian National University (ANU) in Canberra in 1990-91 and at CSIRO Division of Building, Construction and Engineering in Melbourne in 1991-92. He was a lecturer, senior lecturer and an associate professor at the University College, the University of New South Wales (UNSW), the Australian Defense Force Academy (ADFA) in Canberra in 1992-99. Attracted by the English weather, he moved to Birmingham on 1 April 1999 to take up a Chair of Computer Science.
Currently he is the Director of CERCIA (the Centre of Excellence for Research in Computational Intelligence and Applications, http://www.cercia.ac.uk) at the University of Birmingham, UK, a Distinguished Visiting Professor of the University of Science and Technology of China in Hefei, China, and a visiting professor of three other universities. He is an IEEE Fellow and a Distinguished Lecturer of IEEE Computational Intelligence Society. He won the 2001 IEEE Donald G. Fink Prize Paper Award and several other best paper awards. In his spare time, he does the voluntary work as the Editor-in-Chief of IEEE Transactions on Evolutionary Computation, an associate editor or editorial board member of several other journals, and the editor of the World Scientific book series on "Advances in Natural Computation". He has been invited to give more than 45 invited keynote and plenary speeches at conferences and workshops in 16 different countries. He is a Cheung Kong Scholar (Changjian Chair Professor) of the Ministry of Education of China.
His research has been well supported by research councils, government organizations and industry. His major research interests include evolutionary computation, neural network ensembles, and their applications. He has more than 200 refereed technical publications (including 100 journal papers).

Random Field, Network Modularity, Spectral Clustering and Beyond

Hiroshi Mamitsuka, Professor & PhD
Bioinformatics Center, Institute for Chemical Research, Kyoto University, Japan
Email: mami@kuicr.kyoto-u.ac.jp
Personal Website: http://www.bic.kyoto-u.ac.jp/pathway/mami

Abstract: In this presentation, we address the issue of clustering numerical vectors with a network. An example of this setting is clustering genes which can be numerically measured in expression, and at the same time, a gene network can be given from another data source. Web pages can be also numerically vectorized by their contents, e.g. term frequencies, and at the same time, are hyperlinked to each other. Our focus is on the optimal combination of these two heterogeneous data sources. I'll show you two different approaches for this issue. The first approach assumes that a gene network is a random field where random variables take expression values. In contrast to standard random fields such as Markov random fields or Gibbs random fields, we use the idea of network modularity, a network property strongly related with clustering. The second approach directly combines a graph partitioning algorithm with k-means for clustering numerical values, based on the idea of spectral clustering. We examined the performance of the proposed methods using synthetic as well as real-world datasets. The possibility and future of the methods for combining two different data sources will be discussed.

Bio-Sketch: Hiroshi Mamitsuka (http://www.bic.kyoto-u.ac.jp/pathway/mami) is Professor of Bioinformatics Center in Institute for Chemical Research at Kyoto University, where he joined as Associate Professor in 2002 and has worked as Professor since 2005. He is jointly appointed as Professor of School of Pharmaceutical Sciences at Kyoto University since 2006. His prior position from 1991 to 2002 was a research staff member of NEC Research Laboratories in Japan, where he devoted himself to the methodological and empirical researches on machine learning and data mining and developed a wide variety of software programs for real-world problems such as telecom churn management, web usage analysis and product recommendation in E-commerce. Almost at the same time, he was involved with an international research project, called “Real World Computing”, which was sponsored by Japan’s MITI from 1993 to 2001. He received his PhD in Information Sciences from the University of Tokyo in 1999, and M.E. in Information Engineering and B.S. in Biochemistry and Biophysics from the same university in 1991 and 1988, respectively. He has over 60 publications in machine learning, data mining, bioinformatics and biology such as Immunology and molecular evolution. Their focus is mainly on the development of new machine learning techniques and their applications to real issues mainly in biology. A particular emphasis can be placed on that his conference publications include a lot of those appeared in highly competitive conferences. In addition to these research articles, he has 10 patents issued in Japan or in the U.S. He is a co-founder of Japanese Society of Bioinformatics and currently a board member of Asian Association of Societies of Bioinformatics (AASBi). His current research interests are in statistical approaches in data mining and machine learning and applications to bioinformatics.

Using Open Data to Build Intelligent Software

David G. Stork, Chief Scientist & PhD
Ricoh Innovations, USA,
Email: stork@rii.ricoh.com
Personal Website: http://www.rii.ricoh.com/~stork/index.html

Abstract: The building of intelligent software, for speech recognition, language understanding, computer vision, and other applications, has been pursued for decades. While theory and algorithms have been making steady progress, experience has shown that data has become the bottleneck. The size of training data is now the primary factor in an intelligent system’s success. Furthermore, there is a “data gap” between industry and academic research; Industry researchers have access to orders of magnitude more data than academic researchers. This “data gap” is increasingly becoming a “relevancy gap,” where academic research on small data set is becoming less relevant to practitioners.
This presentation will motivate the open acquisition and publication of data. The open acquisition of data over the Web will create data sets even larger than ones currently available to industry researchers. The open publication of data will make data available to all researchers from anywhere around the world. Examples from the Open Mind Initiative and other projects will be discussed.

Bio-Sketch: David G. Stork (http://www.rii.ricoh.com/~stork/index.html) is Chief Scientist of Ricoh Innovations and Visiting Lecturer in Statistics at Stanford University. A graduate in Physics from the Massachusetts Institute of Technology and the University of Maryland, he has held academic appointments at leading colleges and universities in departments and programs in eight academic disciplines: physics, mathematics, computer science, electrical engineering, statistics, neuroscience, psychology, and art and art history. He holds 35 patents and has published over 130 scholarly publications and five books, including Pattern Classification (2nd ed.), the world’s best-selling textbook in the field, Seeing the Light, the world’s best-selling textbook on optics for non-scientists, and HAL’s Legacy: 2001’s computer as dream and reality, which served as the source for his PBS television documentary 2001: HAL’s Legacy. His research interests include pattern classification, machine learning, image analysis of art, concurrency theory, computer speechreading (lipreading) and computational sensing and imaging. He was Artist-in-Residence through the New York Council of the Arts, and has lectured on computer analysis of art at the National Gallery London, Metropolitan Museum of Art, Wadsworth Atheneum, Courtauld Institute, Art Institute of Chicago, and many other leading art institutions. He is the co-editor of the forthcoming proceedings volume, Computer image analysis in the study of art and has delivered over four dozen distinguished or plenary lectures.

Non-Number-Crunching Computation Inspired by Biology: Combinatorics, Graph Theory and Language

Bailin Hao, Professor & Member CAS
T-Life Research Center & Department of Physics, Fudan University in Shanghai, China
Institute of Theoretical Physics, Academia Sinica, Beijing, China
Santa Fe Institute, Santa Fe, New Mexico, USA
Email: hao@itp.ac.cn
Personal Website: http://www.itp.ac.cn/~hao/

Abstract: Biology now generates huge amount of data. Number-crunching computations from finding genes in genomes to simulating biochemical processes in living cells and organisms in physiological and pathological conditions present great challenges for many years to come. In this talk we will touch on some non-number-crunching computation problems inspired by the study of real genomic data. These problems have not reached the height of “Intelligent Computation” yet but require the use of combinatorics, graph theory and formal language thoery. All examples are taken from our own bioinformatics work in recent years.

Bio-Sketch: Bailin Hao (http://www.itp.ac.cn/~hao/) is the Head of T-Life Research Center and Professor of Department of Physics, Fudan University, in Shanghai, China, since 2001. He is currently an External Faculty of Santa Fe Institute. He gradauted from Kharkov State University in Kharkov, Ukraine, in 1959 and worked at the the Institute of Physics, Academia Sinica, in Beijing as Research Asistant (1959-1963), Research Associate (1963-1978) and Research Professor (1978). He became a founding member of the Institute of Theoretical Physics, Academia Sinica in 1978. He served as Deputy-Director of the Institute of Physics (1978) and the Institute of Theoretical Physics (1984-1987), and Director of the Institute of Thoretical Physics (1990-1994). He has been working in statistical physics, computational physics, nonlinear science and theoretical life science. He published more than 150 scientific papers and authored or edited 22 books (8 in English). He was elected a Member of the Chinese Academy of Sciences in November 1980 and a Member of The Third World Academy of Science (TWAS) in 1995. He served as the Chairman of the 19th IUPAP International Conference on Statistical Physics in 1995. He received twice the National Award in Natural Science (1993, 2000) and several awards of CAS. He received the Classic Citation (1981-1998) Award from ISI in 2000 and the Science and Technological Progress Award from the Hong Kong based He-Liang-He-Li Foundation in 2001.