Publicado em: 11/03/2014

Palestras com prof. Peter Gärdenfors da University of Lund – Sweden – no INF

O professor Peter Gärdenfors da University of Lund, Sweden realizará duas palestras no Instituto de Informática. As palestras acontecerão nos dias 17 e 18 de março, intituladas “Conceptual space as a tool for knowledge representation” e “Mind-reading robots”.

Dia: 17/03 – segunda-feira
Horário: 15h30min
Local: Prédio 43413(67) – Sala 106 – INF

TITULO: “Conceptual space as a tool for knowledge representation“

RESUMO: Within computer science, the main methods for representing knowledge have been symbolic or connectionist (using neural networks). As a third option, I have proposed conceptual spaces that are based on geometrical structures in dimensional spaces. In the talk, I will discuss that advantages and disadvantages of the three approaches. I argue that, at least in relation to models of concept formation, conceptual spaces have the greatest explanatory power. I will illustrate the theory with some applications to geoinformatics.

Dia: 18/3 – Terça-feira
Horário: 15h30min
Local: Prédio 43413(67) – Auditório Inferior – INF

TITULO: “Mind-reading robots“

RESUMO: Interaction between humans and robots cannot build only on linguistic communication, but also depends on our ability to understand each other’s goals and desires. In contrast to other animals, humans are good at mind-reading in the sense that they can represent the contents of the minds of others. I shall divide the competence for mind-reading (also known as theory of mind) into five components: representing (1) the emotions, (2) the desires, (3) the attentions, (4) the intentions and (5) the beliefs and knowledge of others. Recent attempts to exploit control theory and other programming techniques for modelling various cognitive functions will be discussed and I will outline how this modelling approach can be combined with the analysis of mind-reading. I then turn to the question of how these techniques can be used to construct mind-reading robots. The focus will be on achieving joint attention and joint intention in human-robot interaction.

BIO: Peter Gärdenfors is a professor of cognitive science at the University of Lund, Sweden. He is a member of the Royal Swedish Academy of Letters, History and Antiquities and recipient of the Gad Rausing Prize. He is member of the Prize Committee for the Prize in Economic Sciences in Memory of Alfred Nobel since 2011. He received his doctorate from Lund University in 1974; his thesis title was “Group Decision Theory”. Internationally, he is one of Sweden’s most notable philosophers and cognitive scientists. Peter has published more than 250 articles in various journals and books including Knowledge in Flux: Modeling the Dynamics of Epistemic States, MIT Press, How Homo Became Sapiens: On the Evolution of Thinking, Oxford University Press, and The Dynamics of Thought, Springer Verlag. Previous research focussed on philosophy of science, decision theory, belief revision and nonmonotonic reasoning. Main current research interests are concept formation (using conceptual spaces based on geometrical and topological models), cognitive semantics, models of knowledge and information and the evolution of cognition.