Ontology Based Information Extraction

EU-IST Integrated Project (IP) IST-2003-506826 SEKT Deliverable D3.5.1(WP3.5) In this document, we propose a framework for reasoning with multi-version ontology, in which a temporal logic is developed to serve as its semantic foundation. We show that the temporal logic approach can provide a solid semantic foundation which can support various requirements on multi-version ontology reasoning. We have implemented the prototype of MORE (Multi-version Ontology REasoner), which is based on the proposed framework. In this document, we present the design of the interfaces for MORE, and discuss its implementation.

Comparative evaluation of Machine Learning (ML) systems used for Information Extraction (IE) has suffered from various inconsistencies in experimental procedures. This paper reports on the results of the Pascal Challenge on Evaluating Machine Learning for Information Extraction, which provides a standardised corpus, set of tasks, and evaluation methodology. The challenge is described and the systems submitted by the ten participants are briefly introduced and their performance is analysed.

Business Intelligence (BI) requires the acquisition and aggregation of key pieces of knowledge from multiple sources in order to provide valuable information to customers or feed statistical BI models and tools. The massive amount of information available to business analysts makes information extraction and other natural language processing tools key enablers for the acquisition and use of that semantic information. We describe the application of ontology-based extraction and merging in the context of a practical e-business application for the EU MUSING Project where the goal is to gather international company intelligence and country/region information. The results of our experiments so far are very promising and we are now in the process of building a complete end-to-end solution.

Knowledge extraction is increasingly attracting the attention of researchers from different disciplines, as a means to automate complex tasks that rely on bulk textual resources. However, the configuration of many devices in the networking field continues to be a labor intensive task, based on the human interpretation and manual entry of commands through a text-based user interface. Typically, these Command-Line Interfaces (CLIs) are both device and vendor-specific, and thus, commands differ syntactically and semantically for each configuration space. Because of this heterogeneity, CLIs always provide a "help" feature-i.e., short command descriptions encoded in natural language-aimed to unveil the semantics of configuration commands for network administrators. In this paper, we exploit this feature with the aim of automating the abstraction of device configurations in heterogeneous settings. In particular, we introduce an Ontology-Based Information Extraction (OBIE) system from the Command-Line Interface of network routers. We also present ORCONF, a domain Ontology for the Router CONFiguration domain, and introduce a semantic relatedness measure that quantifies the degree of interrelation among candidate concepts. The results obtained over the configuration spaces of two widely used network routers demonstrate that this is a promising line of research, with overall percentages of precision and recall of 93%, and 91%, respectively.

The first part of this deliverable presents the Pascal Challenge on evaluating machine learning methods for Information Extraction, the systems that we entered into the challenge and the evaluation results. The second part of deliverable focuses on quantitative evaluation of Ontology-Based Information Extraction (OBIE) and uses the semantically annotated corpus, produced in D2.5.1, in order to evaluate the performance of the system.

In this paper, we describe an approach to populate an existing ontology with instance information present in the natural language text provided as input. An ontology is defined as an explicit conceptualization of a shared domain [18]. This approach starts with a list of relevant domain ontologies created by human experts, and techniques for identifying the most appropriate ontology to be extended with information from a given text. Then we demonstrate heuristics to extract information from the unstructured text and for adding it as structured information to the selected ontology. This identification of the relevant ontology is critical, as it is used in identifying relevant information in the text. We extract information in the form of semantic triples from the text, guided by the concepts in the ontology. We then convert the extracted information about the semantic class instances into Resource Description Framework (RDF3) and append it to the existing domain ontology. This enables...

Much hope has been placed in the past on Educational Television for improving educational standards and distance education. However, in many countries the impact of this technology has been limited. Some of the reasons are discussed along with possible technical solutions. One solution that has been developed at HP Labs India is presented. This involves the use of satellite TV channels, using the MPEG format, as the distribution channel for educational content in digital form. Inexpensive TV equipment, including those that work with Direct To Home (DTH) TV transmissions are used to capture TV programs in digital form and store them on a server. The system is also capable of capturing text content being sent with the program in the Electronic Program Guide format as well as content in various other data formats including PDF. All captured content is stored in a digital library, offering students an on-demand video facility on campus. Easy search facilities are also offered to help find programs of interest.

In this paper, we describe an approach to populate an existing ontology with instance information present
in the natural language text provided as input. An ontology is defined as an explicit conceptualization of a
shared domain [18]. This approach starts with a list of relevant domain ontologies created by human
experts, and techniques for identifying the most appropriate ontology to be extended with information from
a given text. Then we demonstrate heuristics to extract information from the unstructured text and for
adding it as structured information to the selected ontology. This identification of the relevant ontology is
critical, as it is used in identifying relevant information in the text. We extract information in the form of
semantic triples from the text, guided by the concepts in the ontology. We then convert the extracted
information about the semantic class instances into Resource Description Framework (RDF3) and append
it to the existing domain ontology. This enables us to perform more precise semantic queries over the
semantic triple store thus created. We have achieved 95% accuracy of information extraction in our
implementation.

Most of existing ontologies construction tools support construction of ontological relations (e.g., taxonomy, equivalence, etc.) but they do not support construction of domain relations, non-taxonomic conceptual relationships (e.g., causes, caused by, treat, treated by, has-member, contain, material-of, operated-by, controls, etc.). Domain relations are found mainly in text sources. TextOntoEx constructs ontology from natural domain text using semantic pattern-based approach. TextOntoEx is a chain between linguistic analysis and ontology engineering. TextOntoEx analyses natural domain text to extract candidate relations and then maps them into meaning representation to facilitate constructing ontology. The paper explains this approach in more details and discusses some experiments on deriving ontology from natural text.