Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The field of linking fragments from text to entities of a knowledge base is currently at an advanced stage: words and compounds can be identified and disambiguated at high levels of accuracy [1, 2, 8]. Entity linking usually enables to provide further information on the entities, browsing, or recommendation. In this work, we consider the use case of providing a knowledge panel that pops up on mouse over entities that were identified through annotation. The knowledge panel is used to explain the entities identified by the annotator with concise fact-based information units and also enables further browsing (i.e. navigating through related entities). Entity summarization enables to filter all information that is available about an entity (often more than 1000 facts) and to select a small fraction that will be presented to the user.

With this demo, we propose loose coupling between automatic entity linking and entity summarization systems with the “Internationalization Tag Set 2.0” (ITS 2.0) W3C recommendation [3]. We exemplify the feasibility of the lightweight integration with the applications DBpedia Spotlight [1] (as a entity linking tool) and LinkSUM [4] (as an entity summarization tool interfaced via the SUMMA API [5]). Both applications use DBpediaFootnote 1 as a knowledge base.

2 Implementation

We make use of ITS 2.0 by using the its-ta-ident-ref attribute as the combining element between entity linking and entity summaries. This attribute can be used with HTML elements such as <span> tags (for an example see Listing 1). The system is implemented in accordance to the following workflow:

  1. 1.

    Automatically identify and annotate entities of a text.

  2. 2.

    Register a knowledge panel for each identified entity (on mouseover).

figure a
Fig. 1.
figure 1

Automatically annotated excerpt of a Wikipedia article (https://en.wikipedia.org/w/index.php?title=Angela_Merkel&oldid=709980123.) and the summaClient knowledge panel with a summary by LinkSUM.

We had to extend the DBpedia Spotlight jQuery plugin in order to enable ITS 2.0 output.Footnote 2 The system uses a Web service in order to anntoate one or more text paragraphs with entities from the DBpedia knowledge base. The entity summaries are produced by LinkSUM [4]. The produced summaries are displayed as knowledge panels via the summaClientFootnote 3 implementation [5]. The original implementation has been extended by the qSUM method. It registers mouseover events for all elements with an its-ta-ident-ref attribute. The respective knowledge panel is then shown at the position of the respective annotation. The LinkSUM entity summarization system can be easily exchanged by another entity summarization system that implements the SUMMA API.

A screenshot of an automatically annotated text is provided in Fig. 1 and a do-it-yourself example is provided in Listing 2.

figure b

3 Related Work

Our work on ELES was inspired by Denny Vrandečić’s qLabelFootnote 4 project. qLabel is a jQuery-based application that uses text fragments, annotated with its-ta-ident-ref references to Wikidata [7], for translation between different languages. For this, qLabel leverages the wealth of multi-lingual labels available in Wikidata. A Wikidata-based summarization system that uses the SUMMA API can be easily combined with qLabel via qSUM.

There are also a number of proprietary solutions that combine annotation and knowledge panels. The referFootnote 5 application, partly described in [6], supports automatic annotation in combination with knowledge panels. The application is well integrated with additional browsing features and a complete graph panel that can be enabled at the top of each page. Other proprietary solutions include the Bing Knowledge WidgetFootnote 6 and Ontotext’s NowFootnote 7. Most of the proprietary solutions are highly customized and the annotation and knowledge panel parts are often strongly connected.

4 Summary

With ELES, we propose loose coupling between automatic entity linking and entity summarization systems via ITS 2.0. We exemplify the lightweight integration approach with the applications DBpedia Spotlight and the qSUM method of the SUMMA entity summarization interface.