Tutorial Lectures

Lecture 1. Ning Zhong “Research Issues and Challenges on Brain Informatics”

	Ning Zhong, zhong.ning.wici@gmail.com The International WIC Institute/Beijing University of Technology, China, and Maebashi Institute of Technology, Japan

Brain Informatics (BI) is a new interdisciplinary and multidisciplinary field that focuses on studying the mechanisms underlying the human information processing system. It brings together researchers and practitioners from diverse fields to explore the main research problems that lie in the interplay between the studies of human brain and the research of informatics, by using powerful equipment, including functional magnetic resonance imaging (fMRI), electroencephalogram (EEG), positron emission tomography (PET), and eye-tracking as well as various wearable, ubiquitous, active, micro and nano devices. The systematic BI methodology has resulted in the brain big data, including various raw brain data, data-related information, extracted data features, found domain knowledge related to human intelligence, and so forth. In this talk, I demonstrate a systematic approach to an integrated understanding of macroscopic and microscopic level working principles of the brain by means of experimental, computational, and cognitive neuroscience studies, as well as utilizing advanced Web intelligence centric information technologies. I discuss research issues and challenges from three aspects of Brain Informatics studies that deserve closer attention: systematic investigations for complex brain science problems, new information technologies for supporting systematic brain science studies, and Brain Informatics studies based on Web intelligence research needs. These three aspects offer different ways to study traditional cognitive science, neuroscience, mental health and artificial intelligence.

Lecture 2. Boris Mirkin “Data Clustering: Some Topics of Current Interest”

	Boris Mirkin,bmirkin@hse.ru Professor of Department of Data Analysis and Artificial Intelligence of HSE, Leading Research Fellow of International Laboratory of Decision Choice and Analysis http://www.hse.ru/en/org/persons/3954058 Areas of expertise: mathematical methods of cluster analysis and decision making, based on experimental and textual information and applied to sociology, bioinformatics, organizational systems, development of interpreting systems.

As is well known, data cluster is a subset of data entities that are similar to each other and dissimilar from the rest. After a short introduction pointing to different roles clustering plays in machine learning, data mining, and in knowledge discovery, I will move on to describe advances into three issues of current interest: the “right” number of clusters, community detection in networks, and consensus clustering.

The issue of the right number of clusters, in the classical paradigm of looking only at the data structure, has received a considerable boost recently with distinguishing between “after clustering”, “while clustering” and “before clustering” approaches, at least in the clustering framework oriented at minimizing the square error criterion. Moreover, further developments in the digital world lead to availability of various data related to different aspects of the same objects. This allows switching to a better criterion for the number of clusters: that number is right that leads to compatible results on the different aspects.

Community detection in networks, from 2000 on, generated two approaches: modularity criterion and spectral clustering involving the so-called Laplacian normalization of similarity data. Within the least-squares approximation approach both of them can be differently accommodated leading to yet another, although quite natural, “semi-average” clustering criterion.

Consensus clustering is a recent approach to integrating different clustering results. After introducing the basic concepts, I turn to reviewing a recent joint work with A. Shestakov, in which superiority of an implementation of a virtually unknown approach by Mirkin and Muchnik (1981, in Russian), also involving the semi-average criterion, has been demonstrated.

Lecture 3. Luiz F. Autran M. Gomes “Robustness in Discrete Multi-Criteria Decision Analysis”

	Luiz F. Autran M. Gomes, autran@ibmecrj.br Professor of Management in Ibmec University, Rio de Janeiro, Brazil; Researcher of the National Research Council (CNPq) of Brazil; Member of the National Academy of Engineering (Brazil), http://br.viadeo.com/en/profile/luiz-f-autran-m.gomes Areas of expertise: Operations Research, Decision Making, Decision Aids, Organizational Decision Making, Graduate and undergraduate teaching of Multicriteria Decision Aiding.

Multi-Criteria Decision Analysis (MCDA) is the area of the Decision Sciences that focus on structuring, analyzing, modeling, solving, and recommending a solution to decision problems in the presence of multiple criteria. One important issue in MCDA has to do with Robustness Analysis in discrete MCDA. Although there is not a single definition accepted by the scientific community, we can conveniently refer to robustness as the ability of a solution to cope with uncertainties. Different sources of uncertainty interact in a decision problem, some reflecting more or less arbitrary choices of the decision analyst and others concerning external uncertainties. In this sense, it has been proposed that the robustness concern needs to be explicit in a problem such that the robustness analysis is driven by a specific aim. A new robustness analysis framework is proposed where robustness of a solution in a decision aiding process is measured by the distance from that solution to an expected outcome, chosen by the decision-aiding analyst. Therefore, the robustness concern concentrates on changes in criteria weights as well as in trade-off rates, as they are defined in a given MCDA method. Two main contributions are introduced: a local robustness measure, defined in terms of a distance among rankings; and a global robustness measure, as an adaptation of the minimax-regret rule to select a global robust solution.