主题: Information Granularity in Intelligent Data Analysis: Design Studies主讲人: Witold Pedrycz地点: 松江校区2号学院楼226室时间: 2017-03-28 10:00:00组织单位: 信息科学与技术学院
报告人简介：WitoldPedrycz (IEEE Fellow, 1998)is Professor and Canada Research Chair (CRC) in Computational Intelligence inthe Department of Electrical and Computer Engineering, University of Alberta,Edmonton, Canada. He is also with the Systems Research Institute of the PolishAcademy of Sciences, Warsaw, Poland. In 2009 Dr. Pedrycz was elected a foreignmember of the Polish Academy of Sciences. In 2012 he was elected a Fellow ofthe Royal Society of Canada. Witold Pedrycz has been a member of numerousprogram committees of IEEE conferences in the area of fuzzy sets andneurocomputing. In 2007 he received a prestigious Norbert Wiener award from theIEEE Systems, Man, and Cybernetics Society. He is a recipient of the IEEE CanadaComputer Engineering Medal, a Cajastur Prize for Soft Computing from the EuropeanCentre for Soft Computing, a Killam Prize, and a Fuzzy Pioneer Award from theIEEE Computational Intelligence Society.
His main research directions involveComputational Intelligence, fuzzy modeling and Granular Computing, knowledgediscovery and data mining, fuzzy control, pattern recognition, knowledge-basedneural networks, relational computing, and Software Engineering. He haspublished numerous papers in this area. He is also an author of 15 researchmonographs covering various aspects of Computational Intelligence, data mining,and Software Engineering.
Dr. Pedrycz is vigorously involved ineditorial activities. He is an Editor-in-Chief of Information Sciences, Editor-in-Chiefof WIREsData Mining and Knowledge Discovery (Wiley), and Int. J. of Granular Computing(Springer).He serves on an AdvisoryBoard of IEEE Transactions on Fuzzy Systems and is a member of a numberof editorial boards of other international journals.
Information granules play a pivotal role in acquiring, representing,processing, and communicating knowledge at a suitable level of abstraction.Designing information granules is paramount importance to all pursuits ofGranular Computing.
讲座摘要：This presentation offers a comprehensive and systematicallystructured overview of methodologies and algorithms of designing information granulesalong with a suite of representative applications in data analysis anddecision-making. The taxonomy embraces two main categories of data-driven andknowledge-oriented approaches. We introduce and discuss a principle ofjustifiable granularity, which serves as a key design vehicle facilitating aformation of information granules completed on a basis of availableexperimental evidence. Recent advancesof the principle are discussed including (i) a collaborative version of theprinciple supporting data analysis carried out in the presence of distributeddata, (ii) context-based version of the principle incorporating auxiliarysources of knowledge, and (iii) its hierarchical version facilitating handlingexperimental evidence being available at several levels of specificity(abstraction). A collection of design scenarios supporting a formation ofhierarchies of information granules of higher type and higher order ispresented.
In the realm of data analysis, we discuss a collaborative mode ofdiscovery of relationships through constructing granular bi-directional andmulti-directional associative memories and stacked granular autoencoders.