DIVINE
Content providers and analysts increasingly rely on combining multiple data sources to build comprehensive, up-to-date, interlinked information spaces. Techniques for integrating data sources and tracking their evolution are crucial for achieving these goals. DIVINE (Dynamic Integration and Visualization of Information from Multiple Evidence Sources) focuses on developing such techniques, where a lightweight seed ontology acts as the focal point for integrating new evidence derived from multiple, evolving data sources. As such, the project addresses ontology evolution research.
A modular and scalable portfolio of evidence acquisition services integrates structured, unstructured and social sources by crawling public Web documents, querying Linked Open Data repositories, aggregating resource annotations from Web 2.0 applications. Validation processes are triggered for missing or conflicting evidence. Since evidence from third-party sources is inherently uncertain, source-specific confidence value is assigned to each new fact. A spreading activation network utilizes the collected evidence in conjunction with the confidence values for extending the seed ontology.
Domain changes are monitored to derive knowledge evolution patterns. For each ontology element changes over time in confidence values are tracked. Data services and dynamic visualizations reveal rising, declining or cyclic patterns in the recorded values. Such temporal patterns are important indicators which shed light on the evolution of knowledge and on the underlying processes that drive this evolution.
Use cases on news media monitoring and environmental knowledge management demonstrate and evaluate the system’s capabilities for structuring large knowledge repositories, uncovering flows of information between stakeholders, and assessing the reception, understanding and remembering of organization’s external communication.
RAVEN
RAVEN (Relation Analysis and Visualization for Evolving Networks) combines distributed file-level intelligence data with application-level data from enterprise portals to build a comprehensive semantic repository of an organization’s information assets. A social layer is added by enabling users to release non-confidential interpersonal messages to the repository, and by extracting Microformats and RDF annotations from retrieved documents. Third-party resources are added through a Web mining and media monitoring platform.
Within this composite information space RAVEN determines trends in the frequency and semantic orientation of terms and concepts. Ambiguity and subtle incremental change of tonal expressions between different versions of a document complicate sentiment detection, which is addressed by a spreading activation-based approach. The resulting annotations can be used, for example, to track media coverage on a recently launched product line.
Visualization of temporal-semantic relations help analysts comprehend and utilize large enterprise data sets. Developed visualization components are embedded in an AJAX framework consisting of several tightly coupled views providing visual insight into different types of annotations. Scalable, incremental algorithms provide the basis for building semantic repositories and enabling users to visually explore their temporal-semantic relations in real time.
DYONIPOS
The goal of the DYONIPOS project is to optimize and ease the modeling of business processes for organizations by utilizing semantic technologies. In the knowledge management department of the Know-Center a key research area of DYONIPOS is to analyze low-level usage data for identifying the actual task of a knowledge worker and support him with resources and task guides. These analysis and identifications of tasks will serve as a base for automatic process deviation and modeling based on the observed task patterns and information.
In the Call for Proposals 2006 of the FIT-IT Key Research Program Semantic Systems the research proposal DYONIPOS submitted by the Know-Center was honored as the best of the 18 submitted proposals. The DYONIPOS project is carried out in cooperation with the IICM of the Graz University of Technology, HP, and m2n in the next 2 years.
