Catálogo de publicaciones - libros

Text classification is a key technique for handling and organizing text data. The support vector machine(SVM) is shown to be better for the classification among well-known methods. In this paper, the grouping method of the similar words, is proposed for the classification of documents, which is applied to Reuters news and it is shown that the grouping of words has equivalent ability to the Latent Semantic Analysis(LSA) in the classification accuracy. Further, a new combining method is proposed for the classification, which consists of Grouping, LSA followed by the k-Nearest Neighbor classification ( k-NN ). The combining method proposed here, shows the higher accuracy in the classification than the conventional methods of the kNN, and the LSA followed by the kNN. Then, the combining method shows almost same accuracies as SVM.

Object tracking is an important topic in computer vision and image recognition. The probabilistic approach using the particle filter has been recently used for the tracking of moving objects. Based on our trajectory recording system of the soccer scene with multiple video cameras at one view point, we propose the extended approach to increase the tracking robustness and accuracy using the particle filter. The proposed approach makes it possible to pass the necessary particle information using the color histogram and other key factors from one image to the next image, which are taken through the different camera scene with one PC. The performance of the proposed approach is evaluated in the experiments with real video sequence. It is shown that one PC can handle two video images in real-time.

In nature, industry, medicine, social environment, simply everywhere we find a lot of data that bear certain information. A dictionary defines data as facts or figures from which conclusions may be drawn. Data can be classified as either numeric or nonnumeric. The structure and nature of data greatly affects the choice of analysis method. Under the term structure we understand the facts that the data might be not a single number but n-tuples of measurements. Structure is also very closely linked to the reason of data collection and method of measurement. The paper describes the similarities and differences of nature inspired methods and their natural counterparts in light of continuous and discrete properties. Different examples of nature inspired methods are inspected in terms of data, problem domains and inner structure and principles.

The problem of social capital in context of the online social networks is presented in the paper. Not only the specific elements, which characterize the single person and influence the individual’s social capital like static social capital, activity component, and social position, but also the ways of stimulation of the social capital are described.

Kernel Machines, like Support Vector Machines, have been frequently used, with considerable success, in situations in which the input variables are given by real values. Furthermore, the nature of this machine learning algorithm allows extending its applications to deal with other kinds of systems with no vectorial information such as facial images, hand written texts, micro-array gene expressions, or protein chains. The behavior of a number of systems could be better explained if artificial infinite-precision variables were replaced by qualitative variables. Hence, the use of ordinal or interval scales on input variables would allow kernels to be defined for nature-inspired systems directly. In this contribution, two new kernels are designed for applying kernel machines to such systems described by qualitative variables (orders of magnitude or intervals). In addition, the structure of the feature space induced by this kernel is also analyzed.

This study demonstrates the ability of powerful visualization tools (based on the use of connectionist models) to identify network intrusion attempts in an effective and reliable manner. It presents a novel technique to test and evaluate a previously developed network-based intrusion detection system (IDS). This technique applies mutant operators and is intended to test IDSs using numerical data sets. It should be made clear that some mutations were discarded as they did not all provide real life situations. As an application example of the proposed testing model, it has been specially applied to the identification of network scans and mutations of these. The tested Connectionist Agent-Based IDS (CAB-IDS) is used as a method to investigate the traffic which travels along the analysed network, detecting anomalous traffic patterns. The specific tests performed in this study were based on the mutation of one or several variables analysed by CAB-IDS.

We propose in this paper a new kernel, suited for Support Vector Machines learning, which is inspired from the biological world. The kernel is based on Gabor filters that are a good model for the response of the cells in the primary visual cortex and have been shown to be very effective in processing natural images. Furthermore, we build a link between energy-efficiency, which is a driving force in biological processing systems, and good generalization ability of learning machines. This connection can be the starting point for developing new kernel-based learning algorithms.

This paper outlines the abridged history of agent reasoning theories as from the perspective of its implementation inspired by new trends such as ‘teaming’ and ‘learning’. This paper covers how the need for such new notions in agent technology introduced a change in fundamental agent theories and how it can be balanced by inducing some original cognitive notions from the field of ‘artificial mind’. This paper concentrates on the popular agent reasoning notion of Belief Desire Intention (BDI) and outlines the importance of the human-centric agent reasoning model as a step towards the next generation of agents to bridge the gap between human and agent. The current trend including the human-centric nature of agent mind and humanagent teaming is explained, and its needs and characteristics are also explained. This paper reports add-on implementation on BDI in order to facilitate humancentric nature of agent mind. This human-centric nature and concepts such as are utilised to aid human-agent teaming in a simulated environment. The issues in order to make agents more human-like or receptive are outlined.

Trust plays a fundamental role in multi-agent systems in which tasks are delegated or agents must rely on others to perform actions that they themselves cannot do. Dating from the mid-1980s, the Belief-Desire-Intention architecture (BDI) is the longest-standing model of intelligent agency used in multi-agent systems. Part of the attraction of BDI is that it is amenable to logical formalisms such as Wooldridge’s Logic Of Rational Agents (LORA). In a previous paper the present authors introduced a model of trust, here named the Ability-Belief-Commitment-Desire (ABCD) model, that could be implemented within the BDI framework. This paper explores the definition of the ABCD model within the LORA formalism.

This paper describes preliminary work performed to gain an understanding of how to implement collaboration between intelligent agents in a multi-agent system and/or humans. The paper builds on previous research where an agent-development software framework was implemented based on a cognitive hybrid reasoning and learning model. Agent relationships are formed using a three-layer process involving communication, negotiation and trust. Cooperation is a type of relationship that is evident within structured teams when an agent is required to cooperate with and explicitly trust instructions and information received from controlling agents. Collaboration involves the creation of temporary relationships between different agents and/or humans that allows each member to achieve their own goals. Due to the inherent physical separation between humans and agents, the concept of collaboration has been identified as the means of realizing human-agent teams. A preliminary demonstration used to confirm this research is also presented.