Catálogo de publicaciones - libros

Testing is the most common validation method in the soft ware in dustry. It entails the execu tion of the software system in the real envi ron ment. Nevertheless, testing is a cost-in tensive process. Be cause of its conceptual simplicity the combination of formal methods and test methods has been widely advocated. Model checking be longs to the promising candidates for this marriage. The present paper modifies and ex tends the existing approaches in that, after the test case gen eration, a model checking step supports the manual test process. Based on the approach to specifi cation-based construction of test suites, this paper proposes to generate test cases to cover both the specifi cation model and its com ple ment. This helps also to clearly differ enti ate the correct system outputs from the faulty ones as the test cases based on the specifi ca tion are to succeed the test, and the ones based on the complement of the specifica tion are to fail. Thus, the ap proach handles the in an effective manner

In application domains such as medicine, where a large amount of data is gathered, a medical diagnosis and a better understanding of the underlying generating process is an aim. Recordings of temporal data often afford an interpretation of the underlying pattens. This means that for diagnosis purposes a symbolic, i.e. understandable and interpretable representation of the results for physicians, is needed. This paper proposes the use of definitive-clause grammars for the induction of temporal expressions, thereby providing a more powerful framework than context-free grammars. An implementation in Prolog of these grammars is then straightforward. The main idea lies in introducing several abstraction levels, and in using unsupervised neural networks for the pattern discovery process. The results at each level are then used to induce temporal grammatical rules. The approach uses an adaptation of temporal ontological primitives often used in AI-systems.

We propose an event handling mechanism for dealing with spatio-temporal changes. By using this mechanism, we can observe changes of features at diverse viewpoints. We formalize an event that changes a set of features. The relation between events has a hierarchical structure. This structure provides observations of spatial changes at various detailed level.

Debugging, i.e., removing faults from programs, comprises three parts. Fault detection is used to find a misbehavior. Within fault localization the root-cause for the detected misbehavior is searched for. And finally, during repair the responsible parts of the program are replaced by others in order to get rid of the detected misbehavior. In this paper we focus on fault localization which is based on abstract dependencies that are used by the Aspect system [1] for detecting faults. Abstract dependencies are relations between variables of a program. We say that a variable x depends on a variable y iff a new value for y may causes a new value for x. For example, the assignment statement x = y + 1; implies such a dependency relation. Every time we change the value of y the value of x is changed after executing the statement. Another example which leads to the same dependency is the following program fragment:

if ( y < 10) then x = 1; else x = 0;

A new freeway traffic simulator based on a deep model behaviour is proposed. This simulator is defined and developed for helping human traffic operators in taking decisions about predictive control actions in situations prior to congestion. The simulator uses qualitative tags and cognitive events to represent the traffic status and evolution and the temporal knowledge base produced by its execution is very small and it has a high level of cognitive information.

A diagnostic expert system established on model-based reasoning for on-line diagnosis and loss prevention is described in the paper. Its diagnostic ”cause-effect” rules and possible actions (suggestions) are extracted from the results of standard HAZOP analysis. Automatic focusing as well as ”what-if” type reasoning for testing hypothetical actions have been also implemented. The diagnostic system is tested on a granulator drum of a fertilizer plant in a simulation test-bed.

Semantic Web, in order to be effective, needs automatic support for building ontologies, because human effort alone cannot cope with the huge quantity of knowledge today available on the web. We present an algorithm, based on a Machine Learning methodology, that can be used to help knowledge engineers in building up ontologies.

The human eye may present refractive errors as myopia, hypermetropia and astigmatism. This article presents the development of an Ensemble of Classifiers as part of a Refractive Errors Measurement System. The system analyses Hartmann-Shack images from human eyes in order to identify refractive errors, wich are associated to myopia, hypermetropia and astigmatism. The ensemble is composed by three different Machine Learning techniques: Artificial Neural Networks, Support Vector Machines and C4.5 algorithm and has been shown to be able to improve the performance achieved). The most relevant data of these images are extracted using Gabor wavelets transform. Machine learning techniques are then employed to carry out the image analysis.

Extreme Learning Machine, ELM, is a recently available learning algorithm for single layer feedforward neural network. Compared with classical learning algorithms in neural network, e.g. Back Propagation, ELM can achieve better performance with much shorter learning time. In the existing literature, its better performance and comparison with Support Vector Machine, SVM, over regression and general classification problems catch the attention of many researchers. In this paper, the comparison between ELM and SVM over a particular area of classification, i.e. text classification, is conducted. The results of benchmarking experiments with SVM show that for many categories SVM still outperforms ELM. It also suggests that other than accuracy, the indicator combining precision and recall, i.e. value, is a better performance indicator.

In this paper application of kernel based learning algorithms to endoscopy images classification problem is presented. This work is a part the attempts to extend the existing recommendation system (ERS) with image classification facility. The use of a computer-based system could support the doctor when making a diagnosis and help to avoid human subjectivity. We give a brief description of the SVM and LS-SVM algorithms. The algorithms are then used in the problem of recognition of malignant versus benign tumour in gullet. The classification was performed on features based on edge structure and colour. A detailed experimental comparison of classification performance for diferent kernel functions and different combinations of feature vectors was made. The algorithms performed very well in the experiments achieving high percentage of correct predictions.