Logo des Repositoriums

Auflistung nach Autor:in "Selke, Joachim"

1 - 2 von 2
  • Konferenzbeitrag
    Conceptual views for entity-centric search: turning data into meaningful concepts
    (Datenbanksysteme für Business, Technologie und Web (BTW), 2011) Selke, Joachim; Homoceanu, Silviu; Balke, Wolf-Tilo
    The storage, management, and retrieval of entity data has always been among the core applications of database systems. However, since nowadays many people access entity collections over the Web (e.g., when searching for products, people, or events), there is a growing need for integrating unconventional types of data into these systems, most notably entity descriptions in unstructured textual form. Prime examples are product reviews, user ratings, tags, and images. While the storage of this data is well-supported by modern database technology, the means for querying it in semantically meaningful ways remain very limited. Consequently, in entity-centric search suffers from a growing semantic gap between the users' intended queries and the database's schema. In this paper, we introduce the notion of conceptual views, an innovative extension of traditional database views, which aim to uncover those query-relevant concepts that are primarily reflected by unstructured data. We focus on concepts that are vague in nature and cannot be easily extracted by existing technology (e.g., business phone and romantic movie). After discussing different types of concepts and conceptual queries, we present two case studies, which illustrate how meaningful conceptual information can automatically be extracted from existing data, thus enabling the effective handling of vague real-world query concepts.
  • Zeitschriftenartikel
    Information Extraction Meets Crowdsourcing: A Promising Couple
    (Datenbank-Spektrum: Vol. 12, No. 2, 2012) Lofi, Christoph; Selke, Joachim; Balke, Wolf-Tilo
    Recent years brought tremendous advancements in the area of automated information extraction. But still, problem scenarios remain where even state-of-the-art algorithms do not provide a satisfying solution. In these cases, another aspiring recent trend can be exploited to achieve the required extraction quality: explicit crowdsourcing of human intelligence tasks. In this paper, we discuss the synergies between information extraction and crowdsourcing. In particular, we methodically identify and classify the challenges and fallacies that arise when combining both approaches. Furthermore, we argue that for harnessing the full potential of either approach, true hybrid techniques must be considered. To demonstrate this point, we showcase such a hybrid technique, which tightly interweaves information extraction with crowdsourcing and machine learning to vastly surpass the abilities of either technique.