The notion of forgetting, also known as variable elimination, has been investigated extensively in the context of classical logic, but less so in (nonmonotonic) logic programming and nonmonotonic reasoning. The few approaches that exist are based on syntactic modifications of a program at hand. In this paper, we establish a declarative theory of forgetting for disjunctive logic programs under answer set semantics that is fully based on semantic grounds. The suitability of this theory is justified by a number of desirable properties. In particular, one of our results shows that our notion of forgetting can be entirely captured by classical forgetting. We present several algorithms for computing a representation of the result of forgetting, and provide a characterization of the computational complexity of reasoning from a logic program under forgetting. As applications of our approach, we present a fairly general framework for resolving conflicts in inconsistent knowledge bases that are represented by disjunctive logic programs, and we show how the semantics of inheritance logic programs and update logic programs from the literature can be characterized through forgetting. The basic idea of the conflict resolution framework is to weaken the preferences of each agent by forgetting certain knowledge that causes inconsistency. In particular, we show how to use the notion of forgetting to provide an elegant solution for preference elicitation in disjunctive logic programming.

The language of HEX-programs under the answer-set semantics is designed for interoperating with heterogeneous sources via external atoms and for meta-reasoning via higher-order literals in the context of the Semantic Web. As an important technique in managing knowledge bases, the notion of forgetting has received increasing interest in the knowledge-representation area. In this paper, we introduce a semantics-based theory of forgetting for HEX-programs and, in turn, for a class of OWL/RDF ontologies which allows to fully employ semantic information in managing ontologies like editing, merging, aligning, and redundancy removal. 1

Abstract. Traditionally, the quality of ontology matching is measured using precision and recall with respect to a reference mapping. These measures have at least two major drawbacks. First, a mapping with acceptable precision and recall might nevertheless suffer from internal logical problems that hinder a sensible use of the mapping. Second, in practical situations reference mappings are not available. To avoid these drawbacks we introduce quality measures that are based on the notion of mapping incoherence that can be used without a reference mapping. We argue that these measures are a reasonable complement to the well-known measures already used for mapping evaluation. In particular, we show that one of these measures provides a strict upper bound for the precision of a mapping. 1

Traceability links provide support for software engineers in understanding the relations and dependencies among software artifacts created during the software development process. In this research, we focus on re-establishing traceability links between existing source code and documentation to support reverse engineering. We present a novel approach that addresses this issue by creating formal ontological representations for both the documentation and source code artifacts. These representations are then aligned to establish traceability links at the semantic level. Our approach recovers traceability links by utilizing the structural and semantic information in various software artifacts and the linked ontologies are also supported by ontology reasoners to infer implicit relations among these software artifacts.

An ontology can be seen as an explicit description of the concepts and relationships that exist in a domain. In this paper, we address the problem of building an interest ontology and predicting potential friendship relations between users in the social network Live Journal, using features constructed based on the interest ontology. Previous work has shown that the accuracy of predicting friendship links in this network is very low if simply interests common to two users are used as features and no network graph features are considered. Thus, our goal is to organize users ’ interests in an ontology (specifically, a concept hierarchy) and to use the semantics captured by this ontology to improve the performance of learning algorithms at predicting if two users can be friends. We have designed and implemented a hybrid clustering algorithm, which combines hierarchical agglomerative and divisive clustering paradigms, and automatically builds the interest ontology. Furthermore, we have explored the use of this ontology to construct interest-based features and shown that the resulting features improve the performance of various classifiers for predicting friendship links.

With the advent of the Semantic Web, there is an increased availability of meta data (ontologies) that make explicit the semantic commitments associated with data and an urgent need for machine learning algorithms for building predictive models from such data. Usually, there is no unique global interpretation of data from semantically disparate, autonomous sources. Furthermore, it is neither feasible nor desirable to integrate data from sources on the Semantic Web in a centralized data warehouse. In this paper, we formulate the problem of learning classifiers from a set of related, semantically heterogeneous data sources on the Semantic Web from a user’s point of view. We describe a general strategy for transforming algorithms for learning classifiers from data into algorithms for learning classifiers from a set of semantically heterogeneous distributed data sources. We apply this strategy to the task of learning relational Bayesian classifiers from a collection of such data sources. The proposed approach can be generalized to other relational learning algorithms. Our results provide some of the essential elements of approaches for acquiring useful knowledge from information sources that are becoming available on the Semantic Web. 1

Abstract. While it is clear that there are many sources of variation from one development context to another, it is not clear a priori what specific variables will influence the effectiveness of a process in a given context. For this reason, we argue that knowledge about software process must be built from families of studies, in which related studies are run within similar contexts as well as very different ones. Previous papers have discussed how to design related studies so as to document as precisely as possible the values of likely context variables and be able to compare with those observed in new studies. While such a planned approach is important, we argue that an opportunistic approach is also practical. The approach would combine results from multiple individual studies after the fact, enabling recommendations to be made about process effectiveness in context. In this paper, we describe two processes with which we have been working to build empirical knowledge about software development processes: One is a manual and informal approach, which relies on identifying common beliefs or

Abstract. In this paper, we present a formal process model to support the comprehension and maintenance of software systems. The model provides a formal ontological representation that supports the use of reasoning services across different knowledge resources. In the presented approach, we employ our Description Logic knowledge base to support the maintenance process management, as well as detailed analyses among resources, e.g., the traceability between various software artifacts. The resulting unified process model provides users with active guidance in selecting and utilizing these resources that are context-sensitive to a particular comprehension task. We illustrate both, the technical foundation based on our existing SOUND environment, as well as the general objectives and goals of our process model.

Abstract. Traceability links provide support for software engineers in understanding relations and dependencies among software artifacts created during the software development process. In this research, we focus on reestablishing traceability links between existing source code and documentation to support software maintenance. We present a novel approach that addresses this issue by creating formal ontological representations for both documentation and source code artifacts. Our approach recovers traceability links at the semantic level, utilizing structural and semantic information found in various software artifacts. These linked ontologies are supported by ontology reasoners to allow the inference of implicit relations among these software artifacts.

We address the problem of learning predictive models from multiple large, distributed, autonomous, and hence almost invariably semantically disparate, relational data sources from a user’s point of view. We show under fairly general assumptions, how to exploit data sources annotated with relevant meta data in building predictive models (e.g., classifiers) from a collection of distributed relational data sources, without the need for a centralized data warehouse, while offering strong guarantees of exactness of the learned classifiers relative to their centralized relational learning counterparts. We demonstrate an application of the proposed approach in the case of learning link-based Naïve Bayes classifiers and present results of experiments on a text classification task that demonstrate the feasibility of the proposed approach.