20 20 20
Home     Tracks     Dates    Submission    Registration    Venue    Program    Organization    Support   


The 15th International Conference on
Modeling Decisions for Artificial Intelligence
Modelització de Decisions per a la Intel·ligència Artificial
Mallorca, Octubre 15 - 18, 2018
http://www.mdai.cat/mdai2018
Termini de submissió USB Proc.:
DEADLINE: Juny 22nd, 2018

INVITED TALKS


Prof. Jozo Dujmović
San Francisco State University, San Francisco, CA 94132, USA.
Graded Logic Aggregation

Abstract: This paper summarizes basic properties of graded logic – a natural soft computing generalization of classical Boolean logic. Using graded logic aggregators we can build evaluation criteria and apply them in decision engineering. This paper is an extended summary that surveys key concepts of graded logic and graded logic aggregation.


Prof. Michio Sugeno
Emeritus Professor, Tokyo Institute of Technology, Japan
Towards Distorted Statistics based on Choquet Calculus

Abstract: In this study we discuss statistics with distorted probabilities by applying Choquet calculus which we call 'distorted statistics'. To deal with distorted statistics, we consider distorted probability space on the non-negative real line. A (non-additive) distorted probability is derived from an ordinary additive probability by the monotone transformation with a generator. First, we explore some properties of Choquet integrals of non-negative, continuous and differentiable functions with respect to distorted probabilities. Next, we calculate elementary statistics such as the distorted mean and variance of a random variable for exponential and Gamma distributions. In addition, in the case of distorted exponential probability, we define its density function as the derivative of distorted exponential distribution function with respect to distorted Lebesgue measure.

Further, we deal with Choquet calculus of real-valued functions on the real line and explore their basic properties. Then, we consider distorted probability pace on the real line. We also calculate elementary distorted statistics for uniform and normal distributions. Finally, we compare distorted statistics with conventional skew statistics.


Dr. Jordi Nin
BBVA Data \& Analytics, Barcelona, Catalonia, Spain.
Assessing the risk of default propagation in interconnected sectorial financial networks

Abstract: Systemic risk of financial institutions and sectorial companies relies on their inter-dependencies. The inter-connectivity of the financial networks has proven to be crucial to understand the propagation of default, as it plays a central role to assess the impact of single default events in the full system. Here, we take advantage of complex network theory to shed light on the mechanisms behind default propagation. Using real data from the financial company BBVA, we extract the network of client-supplier transactions between more than 140,000 companies, and their economic flows. In this talk , we introduce a basic computational model, inspired by the probabilities of default contagion, that allow us to obtain the main statistics of default diffusion given the network structure at individual and system levels. Achieved results show the exposure of different sectors to the default cascades, therefore allowing for a quantification and ranking of sectors accordingly. As we will show, this information is relevant to propose countermeasures to default propagation in specific scenarios.


Dr. Aida Valls
Universitat Rovira i Virgili, Dept. Computer Engineering and Mathematics, Tarragona
Decision making tools with semantic data to improve tourists' experiences

Abstract: The offices of management of touristic destinations are interested in providing a more user-centered experience that takes into account the personal interests of each visitor or group of visitors. Tourism is a key element of economic wealth of many places, therefore, improving the tourism experience may have a great impact not only on the visitor but also in the place. In this kind of field, the objects of analysis are usually touristic activities (such as parks, museums, events, shopping malls, routes, sports, etc). The amount of options available at each possible destination is usually very large. Their description includes numerical data but also categorical one, sometimes provided as a list of keywords. Exploiting the semantics of those words is crucial to understand the tourist's interests and needs. We will present two decision aiding methods that use domain ontologies to interpret the meaning of the keywords and help the managers and visitors to improve the touristic experience on a certain place.


Dr. Zoe Falomir
Bremen Spatial Cognition Centre, Universität Bremen, Germany.
Improving Spatial Reasoning in Intelligent Systems: Challenges

Abstract: Here we tackle research on spatial thinking when facing two challenges: (i) describing scenes in natural language, and (ii) reasoning about perspectives for object recognition.

Regarding (i) addressing the following research questions is crucial: which kind of spatial features must intelligent systems use? Is location enough? And which kind of reference frames are suitable for communication? Deictic? Relative to the observer? Relative to the object? And what is the most salient object to describe? Intelligent systems must have common grounding with humans so that they can align representations and understand each other. Regarding (ii) addressing the following is decisive: how can we improve spatial perception in intelligent systems so that they can reason about object perspectives? Can tests done to people for measuring their spatial skills be used to model spatial logics?

On one side the challenge is to propose approaches to understand space and communicate about it as humans do. For that, intelligent systems (i.e. robots, tablets) can use computer vision and machine learning algorithms to analyse point clouds, recognise and describe scenes. On the other side the challenge is to propose approaches which solve spatial tests carried out to measure humans' intelligence and to apply these approaches in intelligent systems (i.e. computer games, robots) so that they can improve their spatial thinking, but also help improve humans' spatial thinking by providing them feedback.


Dr. Martí Sánchez-Fibla
University Pompeu Fabra, Barcelona.
Studying the emergence of coordination among embodied learning agents

Abstract: What are the factors influencing joint action and decisions in human cognition that can be modeled and transferred to teams of learning agents? I will present the project INSOCO-DPI2016-80116-P in which we investigate the emergence of collaborative and competitive behaviors in embodied groups of learning agents in spatial game theory scenarios. We study coordination from low-level sensorimotor interaction to higher level decision making. The most relevant results go in the direction of including in the learning information about anticipating the other agent to improve in the realization of a joint task.

We also provide evidence that concepts like Loss-Aversion borrowed from Economics, in which reward is transformed into utility, can accelerate the emergence of cooperation.



 
MDAI 2018

University of Skövde

MDAI - Modeling Decisions

Vicenç Torra, Last modified: 12 : 53 October 09 2018.