OneClass SVM is a popular method for unsupervised anomaly detection. As many other methods, it suffers from the black box problem: it is difficult to justify, in an intuitive and simple manner, why the decision frontier is identifying data points as anomalous or non anomalous. This problem is being widely addressed for supervised models. However, it is still an uncharted area for unsupervised learning. In this paper, we evaluate several rule extraction techniques over OneClass SVM models, while presenting alternative designs for some of those algorithms. Furthermore, we propose algorithms for computing metrics related to eXplainable Artificial Intelligence (XAI) regarding the “comprehensibility”, “representativeness”, “stability” and “diversity” of the extracted rules. We evaluate our proposals with different data sets, including real-world data coming from industry. Consequently, our proposal contributes to extending XAI techniques to unsupervised machine learning models.

In the past few years, several large companies have published ethical principles of Artificial Intelligence (AI). National governments, the European Commission, and inter-governmental organizations have come up with requirements to ensure the good use of AI. However, individual organizations that want to join this effort, are faced with many unsolved questions. This paper proposes guidelines for organizations committed to the responsible use of AI, but lack the required knowledge and experience. The guidelines consist of two parts: i) helping organizations to decide what principles to adopt, and ii) a methodology for implementing the principles in organizational processes. In case of future AI regulation, organizations following this approach will be well-prepared.

The Iuriservice application offers a semantically enabled FAQ search system and case law browser for the Spanish judges in their first appointment. The system is now at the first stage of implementation in the Spanish Judicial School. Users may input questions to the system in natural language to obtain access to a database of experience-based answers to practical day-to-day questions. In order to offer the question-answer pair from the application database that best matches the input question, the search system is enhanced using ontologies and semantic distance calculation. This paper will focus on the description of these technologies, bringing user needs and ontologies to the spotlight, and will show current effectiveness and efficiency results regarding the performance of the FAQ search engine. These results illustrate the enhancement that may be provided by semantic technologies for information retrieval in comparison with other techniques.

The World Wide Web has been revolutionary in terms of impact, scale and outreach. At every level society has been changed in some way by the Web. This Panel will consider likely developments in this extraordinary human construct as we attempt to realise the Next Wave of the Web - a Semantic Web.Nigel Shadbolt will Chair a discussion that will focus on the prospects for the Semantic Web, its likely form and the challenges it faces. Can we achieve the necessary agreements on shared meaning for the Semantic Web? Can we achieve a critical mass of semantically annotated data and content? How are we to trust such content? Do the scientific and commercial drivers really demand a Semantic Web? How will the move to a mobile and ubiquitous Web affect the Semantic Web? How does Web 2.0 relate to the Semantic Web?

This paper shows the ontology development for the second version of the prototype Iuriservice II, an i-FAQ for judicial use containing a repository of professional judicial knowledge. We define the epistemological and ontological levels as separate issues. The domain and middle-out ontology will be connected to the SEKT top and upper level ontology PROTON. In the second part of the paper, we introduce a refinement of the architectural design, especially conceived to improve the user/system interaction by using searching, scoring and matching algorithms for a multidisciplinary ontological processing.

Problem-solving methods are means of describing the inference process of knowledge-based systems. In recent years, a number of these problem-solving methods have been identified that can be reused for building new systems. However, problem-solving methods require specific types of domain knowledge and introduce specific restrictions on the tasks that can be solved by them. These requirements and restrictions are assumptions that play a key role in the reuse of problem-solving methods, in the acquisition of domain knowledge, and in the definition of the problem that can be tackled by knowledge-based systems. In this paper we discuss the different roles assumptions play in the development of knowledge-based systems and provide a survey of assumptions used in diagnostic problem solving. We show how such assumptions introduce targets and bias for goal-driven machine learning and knowledge discovery techniques. © 1998 John Wiley & Sons, Inc.