Keynote Speakers

Data Science and Prediction

Vasant Dhar
Professor and Director, Center for Business Analytics, Stern School of Management, New York University, USA

vdhar

ABSTRACT
Data Science is the study of the generalizable extraction of knowledge from data.[1] A common epistemic requirement in assessing whether new knowledge is actionable for decision making is its predictive power, not just its ability to explain the past. The heterogeneity and scale of data and diversity of analytical methods require data scientists have an integrated skill set, as well as a deep understanding of the craft of problem formulation and the science required to engineer effective solutions. I shall talk about the key issues that arise in industrial strength predictive modeling, including the implications for education in this fast emerging field.

BIO
Vasant Dhar is a data scientist known for his work in machine-learning-based predictive analytics such as the prediction of returns in financial markets, prediction of demand in economic networks, and prediction of outcomes in healthcare. He also researches IT-driven industry transformation, such as the one that is under way in higher education. He writes regularly in the research literature as well as mainstream media including the Financial Times, Wall Street Journal, Forbes, and Wired Magazine. He is editor-in-chief of the Big Data Journal.

[1] Dhar, V., Data Science and Prediction. Communications of the ACM, volume 56, #12, December 2013.

 

 

A Semiotic Approach to Conceptual Modelling

Antonio Furtado
Peter Chen Award, Departamento de Informàtica, Pontifícia Universidade Católica do Rio de Janeiro, Brazil

alf_photograph

ABSTRACT
The work on Conceptual Modelling performed by our group at PUC-Rio is surveyed, covering four mutually dependent research topics. Regarding databases as a component of information systems, we extended the scope of the Entity-Relationship model, so as to encompass facts, events and agents in a three-schemata specification method employing a logic programming formalism. Next we proceeded to render the specifications executable, by utilizing backward-chaining planners to satisfy the agents’ goals through sequences of fact-modification events. Thanks to the adoption of this plan-recognition / plan-generation paradigm, it became possible to treat both business-oriented and fictional narrative genres. To guide our conceptual modelling approach, we identified four semiotic relations, associated with the four master tropes that have been claimed to provide a system to fully grasp the world conceptually.

BIO
Antonio L. Furtado is Professor Emeritus at the Departamento de Informática of the Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio). He obtained his Ph.D. degree in Computer Science from the University of Toronto in 1974. His main areas of interest are Conceptual Modelling of Information Systems, Logic Programming, and Digital Storytelling. Together with Clesio Saraiva dos Santos (Universidade Federal do Rio Grande do Sul) and Erich J. Neuhold (Darmstadt University of Technology), he participated in the very first Entity-Relationship conference (ER 1979, Los Angeles, California).

Since then, he has been publishing his work in journals and technical events, including several participations in ER conferences. In co-authorship with Erich J. Neuhold and with the collaboration of his colleagues Marco A. Casanova and Paulo A. S. Veloso, he wrote the book Formal Techniques for Data Base Design, Berlin: Springer-Verlag, 1986. He also does research on Medieval Literature, having contributed, among other studies, the chapter “The Crusaders’ Grail” to the book The Grail, the Quest and the World of Arthur, organized by Norris J. Lacy (Pennsylvania State University), Cambridge: D. S. Brewer, 2008.

With respect to Conceptual Modelling, he has been investigating, together with colleagues and students, a multi-disciplinary approach that encompasses static, dynamic and behavioural aspects of Information Systems. This approach borrows plan-generation and plan-recognition methods from Artificial Intelligence, as well as several notions from Literary Theory, Linguistics and Semiotics, such as narrative plots, rhetorical tropes and analogy. Besides its application to business domains, it has been used to characterize narrative genres, as part of a Digital Storytelling project.

 

 

Ontological Patterns, Anti-Patterns and Pattern Languages for Next-Generation Conceptual Modeling

Giancarlo Guizzardi
Ontology and Conceptual Modeling Research Group, UFES Brazil and University of Trento, Italy

foto-2 copy

ABSTRACT
In his ACM Turing Award Lecture entitled “The Humble Programmer”, E. W. Dijkstra discusses the sheer complexity one has to deal with when programming large computer systems. His article represented an open call for an acknowledgement of the complexity at hand and for the need of more sophisticated techniques to master this complexity. This talk advocates the view that we are now in an analogous situation with respect to conceptual modeling. We will experience an increasing demand for building Reference Conceptual Models in subject domains in reality, as well as employing them to address classes of problems, for which sophisticated ontological distinctions are demanded.

One of these key problems is Semantic Interoperability. Effective semantic interoperability requires an alignment between worldviews or, to put it more accurately, it requires the precise understanding of the relation between the (inevitable) ontological commitments assumed by different conceptual models and the systems based on them (including sociotechnical systems). This talk advocates the view that an approach that neglects true ontological distinctions (i.e., Ontology in the philosophical sense) cannot meet these requirements. The talk discusses the importance of foundational axiomatic theories and principles in the design of conceptual modeling languages and models. Moreover, it discusses the role played by three types of complexity management tools: Ontological Design Patterns (ODPs) as methodological mechanisms for encoding these ontological theories; Ontology Pattern Languages (OPLs) as systems of representation that take ODPs as higher-granularity modeling primitives; and Ontological Anti-Patterns (OAPs) as structures that can be used to systematically identify possible deviations between the set of valid state of affairs admitted by a model (the actual ontological commitment) and the set of state of affairs actually intended by the stakeholders (the intended ontological commitment).

Finally, the talk elaborates on the need for proper computational tools to support a process of pattern-based conceptual model creation, analysis, transformation and validation (via model simulation).

BIO
Giancarlo Guizzardi obtained a PhD (with the highest distinction) from the University of Twente, in The Netherlands. He is currently a visiting professor at the University of Trento (Italy) and an associate researcher at the Laboratory of Applied Ontology (LOA), Institute of Cognitive Sciences and Technology (ISTC), also located in Trento. He is on an extended sabbatical leave from the Computer Science Department of the Federal University of Espírito Santo (UFES), where he leads the Ontology and Conceptual Modeling Research Group (NEMO). He has been working for the past 18 years in the areas of Ontology and Conceptual Modeling. He is the author of nearly 160 publications in these areas, including recipients of best paper awards at conferences such as CAISE and EDOC. He is a former member of the Executive Council and currently a member of the Advisory Board of the International Association for Ontologies and its Applications (IAOA). Over the years, he has been involved in the editorial board of several journals (including Applied Ontology, Semantic Web, Requirements Engineering) and played active roles (PC Chair, General Organizational Chair, Program Board Member) in several international conferences (CAISE, FOIS, EDOC). Finally, his experience in ontology-driven conceptual modeling has also been acquired in a number of industrial projects in domains such as off-shore software development, petroleum and gas, digital journalism, government, telecommunications, product recommendation, and complex media management.