Knowledge Graph Completion seeks to find missing elements in a Knowledge Graph, usually edges representing some relation between two concepts. One possible way to do this is to find paths between two nodes that indicate the presence of a missing edge. This can be achieved through Reinforcement Learning, by training an agent that learns how to navigate through the graph, starting at a node with a missing edge and identifying what edge among the available ones at each step is more promising in order to reach the target of the missing edge. While some approaches have been proposed to this effect, their reward functions only take into account whether the target node was reached or not, and only apply a single Reinforcement Learning algorithm. In this regard, we present a new family of reward functions based on node embeddings and structural distance, leveraging additional information related to semantic similarity and removing the need to reach the target node to obtain a measure of the benefits of an action. Our experimental results show that these functions, as well as the novel use of more modern Reinforcement Learning techniques, are able to obtain better results than the existing strategies in the literature.

Defeasible argumentation is typical of legal and scientific reasoning. A defeasible argument is one in which the conclusion can be accepted tentatively in relation with the evidence known so far, but may need to be retracted as new evidence comes in. This paper analyses the role of defeasible argumentation in the explanation and evaluation of architectural decisions. We analyse technical explanations offered by engineers at Twitter and eBay about several architectural decisions adopted in those systems. We generalize these examples in four argumentation schemes. We also study the typical case of reasoning with a formal model of an architecture, and we infer a fifth argumentation scheme. Finally, we apply Hastings' method of attaching a set of critical questions to each scheme. We show that the existence of critical questions reveals that the inferred schemes are defeasible: in argumentation theory, if a respondent asks one of the critical questions matching a scheme and the proponent of an argument fails to offer an adequate answer, the argument defaults and the conclusion is retracted. This dialogical structure is the basis of typical architectural evaluations. We conclude that the provided evidence supports the hypothesis that defeasible argumentation is employed in architectural evaluation. In this context, a rich catalogue of argumentation schemes is a useful tool for the architect to organize his or her reasoning, critical questions assist the architect in identifying the weak points of his or her explanations, and provide the evaluation team with a checklist of issues to be raised.

One the most challenging problems for enterprise information integration is to deal with heterogeneous information sources on the Web. The reason is that they usually provide information that is in human-readable form only, which makes it difficult for a software agent to understand it. Current solutions build on the idea of annotating the information with semantics. If the information is unstructured, proposals such as S-CREAM, MnM, or Armadillo may be effective enough since they rely on using natural language processing techniques; furthermore, their accuracy can be improved by using redundant information on the Web, as C-PANKOW has proved recently. If the information is structured and closely related to a back-end database, deep annotation ranges among the most effective proposals, but it requires the information providers to modify their applications; if deep annotation is not applicable, the easiest solution consists of using a wrapper and transforming its output into annotations. In this paper, we prove that this transformation can be automated by means of an efficient, domain-independent algorithm. To the best of our knowledge, this is the first attempt to devise and formalize such a systematic, general solution

Software solutions to automate the procurement of web services are gaining importance when technology evolves, the number of providers increases and the needs of the clients become more complex. There are several proposals in this field, but they all have important drawbacks, namely: many of them are not able to check offers and demands for internal consistency; selecting the best offer usually relies on evaluating linear objective functions, which is quite a naive solution; the language to express offers is usually less expressive than the language to express demands; and, last but not least, providers cannot impose constraints on their clients. In this article, we present a solution to overcome these problems that relies on constraint programming; furthermore, we present a run-time framework, some experimental results, and a comparison with other proposals.

Abstract Multiparty interactions are a powerful mechanism for coordinating several entities that need to cooperate in order to achieve a common goal. In this paper, we present an algorithm for implementing them that improves on previous results in that it does not require the whole set of entities or interactions to be known at compile‐ or run‐time, and it can deal with both terminating and non‐terminating systems. We also present a comprehensive simulation analysis that shows how sensitive to changes our algorithm is, and compare the results with well‐known proposals by other authors. This study proves that our algorithm still performs comparably to other proposals in which the set of entities and interactions is known beforehand, but outperforms them in some situations that are clearly identified. In addition, these results prove that our algorithm can be combined with a technique called synchrony loosening without having an effect on efficiency. Copyright © 2004 John Wiley & Sons, Ltd.

The semantic Web shall enable Web agents an efficient, precise, and comprehensive extraction of knowledge. Nevertheless, this new Web is not likely to be adopted in the immediate future. We present a specification of a new framework in order to extract knowledge from today's dynamics nonsemantic Web. Our proposal is novel in that it associates semantics with the information extracted, which improves agent interoperability; it can also deal with changes to the structure of a Web page, which improves adaptability; furthermore, it achieves to delegate the knowledge extraction procedure to specialist agents, easing software development and promoting software reuse and maintainability.

Abstract In the context of the Multiparty Interaction Model, fairness is used to insure that an interaction that is enabled sufficiently often in a concurrent program will eventually be selected for execution. Unfortunately, this notion does not take conspiracies into account, i.e. situations in which an interaction never becomes enabled because of an unfortunate interleaving of independent actions; furthermore, eventual execution is usually too weak for practical purposes since this concept can only be used in the context of infinite executions. In this article, we present a new fairness notion, k‐conspiracy‐free fairness, that improves on others because it takes finite executions into account, alleviates conspiracies that are not inherent to a program, and k may be set a priori to control its goodness to address the above‐mentioned problems. Copyright © 2003 John Wiley & Sons, Ltd.

This paper describes an approach based on evolutionary algorithms, hierarchical decision rules (HIDER), for learning rules in continuous and discrete domains. The algorithm produces a hierarchical set of rules, that is, the rules are sequentially obtained and must therefore be tried until one is found whose conditions are satisfied. Thus, the number of rules may be reduced because the rules could be inside of one another. The evolutionary algorithm uses both real and binary coding for the individuals of the population. We tested our system on real data from the UCI repository, and the results of a ten-fold cross-validation are compared to C4.5s, C4.5Rules, See5s, and See5Rules. The experiments show that HIDER works well in practice.

We present a light-weight approach for the automatic verification of requirements. This approach is not based on natural language parsing techniques but on the representation of requirements in XML. In our approach, XSLT stylesheets are used not only to automatically generate requirements documents, but also to provide verification-oriented heuristics as well as to measure the quality of requirements using some verification-oriented metrics. These ideas have been implemented in REM, an experimental XML-based requirements management tool also described.

This article reports on an error-repair algorithm for LR parsers. It locally inserts, deletes or shifts symbols at the positions where errors are detected, thus modifying the right context in order to resume parsing on a valid piece of input. This method improves on others in that it does not require the user to provide additional information about the repair process, it does not require precalculation of auxiliary tables, and it can be easily integrated into existing LR parser generators. A Yacc-based implementation is presented along with some experimental results and comparisons with other well-known methods.

Separation of concerns has been presented as a promising tool to tackle the design of complex systems in which cross-cutting properties that do no fit into the scope of a class must be satisfied. In this paper, we show that interaction amongst a number of objects can also be described separately from functionality by means of the CAL language, and present a framework that provides the needed infrastructure. It is innovative because it supports open multiparty interactions.