Angela Fogarolli

Wikipedia as a Source of Ontological Knowledge: State of the Art and Application

This chapter motivates that Wikipedia can be used as a source of knowledge for creating semantic enabled applications, and consists of two parts. First, we provide an overview over different research fields which attempt to extract knowledge encoded by humans inside Wikipedia. The extracted knowledge can then be used for creating a new generation of intelligent applications based on the collaborative character of Wikipedia, rather than on domain ontologies which require the intervention of knowledge engineers and domain experts. Second, as a proof of concept, we describe an application whose intelligent behavior is achieved by using Wikipedia knowledge for automatic annotation and representation of multimedia presentations.

Intelligent Mining and Indexing of Multi-language e-Learning Material

In this paper we describe a method to automatically discover important concepts and their relationships in e-Lecture material. The discovered knowledge is used to display semantic aware categorizations and query suggestions for facilitating navigation inside an unstructured multimedia repository of e-Lectures. We report about an implemented approach for dealing with learning materials referring to the same event in different languages. The information acquired from the speech is combined with the documents such as presentation slides, which are temporally synchronized with the video for creating new knowledge through a mapping with a taxonomy representation such as Wikipedia.

Entity popularity on the web: correlating ANSA news and AOL search

Annotated Web content, digital libraries, news and media portals, e-commerce web sites, online catalogs, RDF/OWL knowledge bases and online encyclopedias can be considered containers of named entities such as organizations, persons, locations. Entities are mostly implicitly mentioned in texts or multi-media content, but increasingly explicit in structured annotations such as the ones provided by the Semantic Web. Today, as a result of different research projects and commercial initiatives, systems deal with massive amounts of data that are either explicitly or implicitly related to entities, which have to managed in an efficient way. This paper contributes to Web Science by attempting to measure and interpret trends of entity popularity on the WWW, taking into consideration the occurrence of named entities in a large news corpus, and correlating these findings with analysis results on how entities are searched for, based on a large search engine query log. The study shows that entity popularity follows well-known trends, which can be of interest for several aspects in the development of services and applications on the WWW that deal with larger amounts of data about (named) entities.

Digital Libraies as Support for e-Learning Activities

In e-Learning environments, very often a large number of multimedia learning materials are provided and used in combination. Accessing a repository of unstructured multimedia material for learning purposes is still an open problem. In this paper we present the evaluation process and results of a two evaluation session run in a two year period for accessing the efficacy of a digital library in the learning process. The number of participants was 109. Results show that using such a tool in an e-learning environment can shorten the students learning time, while achieving similar performance.

Discovering Semantics in Multimedia Content Using Wikipedia

Semantic-based information retrieval is an area of ongoing work. In this paper we present a solution for giving semantic support to multimedia content information retrieval in an e-Learning environment where very often a large number of multimedia objects and information sources are used in combination. Semantic support is given through intelligent use of Wikipedia in combination with statistical Information Extraction techniques.

Linking data for public administrations

Private and public enterprises suffer from the exponential growth of new data, becoming available both internally within their organizations as externally on the web. Existing databases and applications are not set up to be using these increasingly complex data sources. We present a novel approach designed for a tax agency for gaining much more powerful and sustainable access to their data sources by linking them internally as well externally with the Web of Data. The flexible and rapid linking of our approach avoids costly mergers or consolidation of databases or in-depth re-programming of data applications. In this paper we describe a prototype which effectively links organizational data about persons, organizations, buildings and locations that are spread over different databases, which enables the inclusion and the connection with the Web of Data.

Entity re-use in a large news corpus

In Digital content,named entities such as organizations, persons, locations are mantioned implicity or explicity using structured annotations. In this paper we attempt to measure an aspect of entity popularity by taking into consideration the occurrences of named entities in 20 years of New York Times articles. This study shows that entity re-use is an interest factor to take into account in the development of entity-centric applications and services.

Domain Independent Semantic Representation of Multimedia Presentations

This paper describes a domain independent approach for semantically annotating and representing multimedia presentations. It uses a combination of techniques to automatically discover the content of the media and, though supervised or unsupervised methods, it can generate a RDF description out of it. The domain independence is achieved using Wikipedia as a source of knowledge instead of domain Ontologies. The described approach can be relevant for understanding multimedia content which can be used in Information Retrieval, categorization and summarization.