Catálogo de publicaciones - libros

Choosing among options and selecting the best alternative is a fundamental component of human decision-making. The best alternative is the result of a mental process called . The evaluated system can be any collection of interrelated components. The system as a whole and its components are expected to have some desired features and satisfy specific requirements. Consequently, system evaluation is a process of determining the extent to which a system satisfies a given set of requirements.

The human brain consists of a neural system as hardware and a language system as software. It is, therefore, possible to take two approaches to create the human brain. While the hardware-centered approach is based on computational neuroscience, it is possible to base the software-centered approach on linguistics.

Brain-style computing is considered as one of the main research areas in creating the brain. We take a language-based approach to brain-style computing. To this aim, we have adopted as the basic theory Systemic Functional Linguistics (SFL) initiated by Halliday.

Following Halliday’s four principles in the design of human language, we have implemented the computational model of language in context, called the Semiotic Base, and we have developed a set of algorithms of text understanding and generation using this model. The language used in this study is Japanese.

As an application of the models, we are developing Brain-Style Computing System under which we can manage and execute all kinds of computing through meanings. The idea is to verbalize computers by constructing linguistic models of software and hardware applications. Brain-Style Computing System consists of Everyday Language Interface with a Secretary Agent, Semiotic Base, Language Applications, Language Communication Protocol and Language Operating System.

In this talk, I shall discuss some linguistic issues in creating the brain. There are three higher-order functions of the brain concerned with language: processing, utilizing, and learning language. Processing language such as understanding and generation is a basic function with the internal models of language itself and its processing. SFL could reveal what the internal models must be like. SFL could also play an essential role in elucidating the brain functions of language such as thinking with language and learning language.

I shall also show some clinical evidence obtained from studies on aphasia which support the SFL perspective on the system of language. I shall also refer to the brain internal models for motor control and some learning mechanisms in the brain which might be related with language functions.

Case-based reasoning is a paradigm in machine learning whose idea is that a new problem can be solved by noticing its similarity to a set of problems previously solved. We propose a new approach to case-based reasoning. It is based on rough set theory that is a mathematical theory for reasoning about data. More precisely, we adopt Dominance-based Rough Set Approach (DRSA) that is particularly appropriate in this context for its ability of handling monotonicity relationship between ordinal properties of data related to monotonic relationships between attribute values in the considered data set. In general terms, monotonicity concerns relationship between different aspects of a phenomenon described by data: for example, “the larger the house, the higher its price” or “the closer the house to the city centre, the higher its price”. In the perspective of case-based reasoning, we propose to consider monotonicity of the type “the more similar is to , the more credible is that belongs to the same set as ”. We show that rough approximations and decision rules induced from these approximations can be redefined in this context and that they satisfy the same fundamental properties of classical rough set theory.

The scientific research in the area of computational trust and reputation mechanisms for virtual societies is a recent discipline oriented to increase the reliability and performance of electronic communities by introducing in such communities these well known human social control mechanisms.

Many realistic decision aid problems are fraught with facets of ambiguity, uncertainty and conflict, which hamper the effectiveness of conventional and fuzzy preference modeling approaches, and command the use of more expressive representations. In the past, some authors have already identified Ginsberg’s/Fitting’s theory of bilattices as a naturally attractive candidate framework for representing uncertain and potentially conflicting preferences, yet none of the existing approaches addresses the real expressive power of bilattices, which lies hidden in their associated truth and knowledge orders. As a consequence, these approaches have to incorporate additional conventions and ‘tricks’ into their modus operandi, making the results unintuitive and/or tedious. By contrast, the aim of this paper is to demonstrate the potential of (rectangular) bilattices in encoding not just the problem statement, but also its generic solution strategy.

Multiperson decision making problems involve using the preferences of some experts about a set of alternatives in order to find the best of those alternatives. However, sometimes experts cannot give all the information that they are required. Particularly, when dealing with fuzzy preference relations they can avoid giving some of the preference values of the relation. In the literature these incomplete information situations have been faced giving procedures which are able to compute missing information from the preference relations. However, these approaches usually need at least a piece of information about every alternative in the problem. In this paper, several strategies to manage situations, that is, situations where an expert does not provide information on at least one alternative are presented, and their advantages and disadvantages analised.

Negotiation engines are major components of autonomous agents, because negotiation is one of the most important types of agent interaction. Thus far, most negotiation engines rely on analytic techniques to maximize the social welfare of agent communities. Such engines are developed with total disregard of the possibility of enabling agents to analyze offers made by their negotiation opponents. The analysis of the offers leads to making tradeoffs that result into agreeing on (selecting) solution options that maximize the social welfare of the negotiation agents. Therefore, this paper presents an agent negotiation engine that supports the following: evaluation of solution options, analysis of tradeoffs, analysis of offers, and management of negotiation deadlocks. Moreover, the paper presents a simulation experiment that illustrates the capabilities of the negotiation engine.

We discuss a decision theoretic approach to learn causal Bayesian networks from observational data and experiments. We use the information of observational data to learn a completed partially directed acyclic graph using a structure learning technique and try to discover the directions of the remaining edges by means of experiment. We will show that our approach allows to learn a causal Bayesian network optimally with relation to a number of decision criteria. Our method allows the possibility to assign costs to each experiment and each measurement. We introduce an algorithm that allows to actively add results of experiments so that arcs can be directed during learning. A numerical example is given as demonstration of the techniques.

The aim of this work is to study some of the differences between the additive and multiplicative representations associated with the Analytic Hierarchy Process. We present the Method of Pair Comparisons with the study of its properties from the point of view of representational measurement theory and scaling theory. From the first point of view it is impossible to differentiate two types of representations and therefore, the distinction has to be done in terms of the type of task that the subjects perform. The conclusion establishes some differences and relationships between the task of making judgments of proportion and judgments of distance.

Strategic planning can be schematised as a decision making process where, given a general outline of the desirable future, the decision makers need to choose a set of actions that should coherently lead a system (corporation, institution, city, region, etc.) toward that future. A more sophisticated case is when rather than only choosing actions, the decision maker also decides the allocation of available resources among different available actions. We show that in most cases the problem can be faced using a particular Decision Network with multiple objectives, in which actions are applied simultaneously and are modelled by variables representing the efforts spent on them. The main advantage of the proposed Simultaneous Decision Network is that it can be easily built by a panel of domain experts, under the assumption of the noisy-OR causal interaction. The problem of finding the best strategy in terms of resource allocation is formulated as a combinatorial optimisation, and solved through a multi-objective meta heuristic approach.