The exponential growth of digital information available in companies and on the Web creates the need for search tools that can respond to the most sophisticated information needs. Many user tasks would be simplified if Search Engines would support typed search, and return entities instead of just Web documents. For example, an executive who tries to solve a problem needs to find people in the company who are knowledgeable about a certain topic.In the first part of the book, we propose a model for expert finding based on the well-consolidated vector space model for Information Retrieval and investigate its effectiveness.In the second part of the book, we investigate different methods based on Semantic Web and Natural Language Processing techniques for ranking entities of different types both in Wikipedia and, generally, on the Web.In the third part of this thesis, we study the problem of Entity Retrieval for news applications and the importance of the news trail history (i.e., past related articles) to determine the relevant entities in current articles.We also study opinion evolution about entities: We propose a method for automatically extracting the public opinion about political candidates from the blogosphere.

With the advent of the Semantic Web and Semantic Technologies, ontologies have become one of the most prominent paradigms for knowledge representa-tion and reasoning. However, recent progress in the field faces a lack of well-structured ontologies with large amounts of instance data due to the fact that engineering such ontologies requires a considerable investment of resources. Nowadays, knowledge bases often provide large volumes of data without so-phisticated schemata. Hence, methods for automated schema acquisition and maintenance are sought. Schema acquisition is closely related to solving typi-cal classification problems in machine learning, e.g. the detection of chemical compounds causing cancer. In this work, we investigate both, the underlying machine learning techniques and their application to knowledge acquisition in the Semantic Web.

The general objective of the thesis is to provide domain independent, and resource independent methods and tools for speeding up the ontology development process and is achieved by reusing and re-engineering as much as possible available non-ontological resources (NORs). To fulfil this overall goal, we have decomposed it into the following methodological and technological objectives: The definition of methodological aspects related to the reuse of non-ontolo-gical resources for building ontologies. The definition of methodological aspects related to the re-engineering of non-ontological resources for building ontologies. The creation of a library of patterns for re-engineering non-ontological resources into ontologies. The development of a software library, NOR2O, that implements the suggestions given by the re-engineering patterns.

Mathematics is becoming increasingly collaborative, but software does not sufficiently support that: Social Web applications do not currently make mathematical knowledge accessible to automated agents that have a deeper understanding of mathematical structures. Such agents exist but focus on individual research tasks, such as authoring, publishing, peer-review, or verification, instead of complex collaboration workflows. This work effectively enables their integration by bridging the document-oriented perspective of mathematical authoring and publishing, and the network perspective of threaded discussions and Web information retrieval. This is achieved by giving existing representations of mathematical and relevant related knowledge about applications, projects and people a common Semantic Web foundation. Service integration is addressed from the two perspectives of enriching published documents by embedding assistive services, and translating between different knowledge representations inside knowledge bases. A usability evaluation of a semantic wiki that coherently integrates knowledge production and consumption services points out the remaining challenges in making such heterogeneously integrated environments support realistic workflows. The results of this thesis will soon also enable collaborative acquisition of new mathematical knowledge, as well as the contribution of existing knowledge collections to the Web of Data.

Ontological modelling today is applied in many areas of science and technology, including the Semantic Web. The W3C standard OWL defines one of the most important ontology languages based on the semantics of description logics. An alternative is to use rule languages in knowledge modelling, as proposed in the W3C’s RIF standard. So far, it has often been unclear how to combine both technologies without sacrificing essential computational properties.This book explains this problem and presents new solutions that have recently been proposed. Extensive introductory chapters provide the necessary background for understanding the goals and challenges of this field, whereas advanced chapters discuss novel solutions in full detail. Enriched knowledge representation languages that are introduced include DL Rules, Horn description logics, and DL+safe Rules. In each of these cases, emphasis is put on finding a favourable trade-off between expressiveness and computational complexity. This naturally leads to the light-weight DL rule language ELP which illustrates that expressive ontological modelling and tractable inferencing can indeed go together. Comprehensive references for further reading are provided throughout the book.

The development of knowledge-based systems is usually approached through the combined skills of knowledge engineers (KEs) and subject matter experts (SMEs). One of the most critical steps in this activity aims at transferring knowledge from SMEs to formal, machine-readable representations, which allow systems to reason with such knowledge. However, this is a costly and error prone task. Alleviating the knowledge acquisition bottleneck requires enabling SMEs with the means to produce the desired knowledge representations without the help of KEs. This is especially difficult in the case of complex knowledge types, like processes. The analysis of different application domains uncovers that process knowledge is one of the most frequent knowledge types, whose complexity requires specific means to enable SMEs to represent processes in a computational form. Additionally, such complexity and the increasingly large amount of data that process executions generate in knowledge-intensive domains, like Biology or Astronomy, requires analytical means with high abstraction capabilities to support SMEs in the analysis of such processes. This book presents methods and tools that enable SMEs to acquire process knowledge from the domains, formally represent such knowledge, reason about it, and understand process executions by analyzing their provenance. We describe the utilization of Problem Solving Methods as the main knowledge artifacts for process acquisition and analysis in two innovative ways. First, as formalizations of the reasoning strategies needed for processes and, second, as high-level, domain-independent, and reusable abstractions of process knowledge to provide SMEs with interpretations of process executions.