PowerMagpie—a Semantic ‘Web Browser’

‘Semantic Web’ Browsing & Semantic ‘Web Browsing’

The Semantic Web—as a web of data—can be viewed as a rich, multidimensional hypertext system, on which browsing—as ‘regular’ hypertext browsing or faceted navigation—can be performed.

‘Semantic web’ browsing is about exploring the context of semantic data within semantic data, while semantic ‘web browsing’ is about exploring the semantic context of terms present within a regular web resource and subsequently is about exploring the resource's domain semantics.

PowerMagpie

The primary goal of PowerMagpie is to bring semantic interpretation to web browsing, without involving important effort from the user. This suppose to automatically relate the major terms present in the text of the web page to semantic entities, dynamically discovered in ontologies available on the (Semantic) Web. In order to achieve this goal, PowerMagpie has to handle four major tasks:

1. Identifying relevant terms in the currently browse web page. Not all the terms occurring in the text of the web page have the same importance. It is crucial to identify the ones that characterize the domain covered by the document, and to present first the one that are the most important for the interpretation and understanding of the page.

2. Selecting online ontologies to interpret the domain terms. This task constitutes one the major difference between Power Magpie and its predecessor Magpie. Magpie required for the user to manually select the appropriate ontology for semantic browsing. This was an important limitation since the user may not be familiar with ontologies and had to be aware of the existence of the appropriate ontology, matching the domain of the currently browsed document. This process had to be carried out every time the user would browse a new page about a domain not covered by the selected ontology. In addition, Magpie was bounded by the use of a single ontology. Web documents often relate to several different domains, requiring the knowledge contained in several different ontologies to be interpreted. In PowerMagpie, the goal is to automate the selection of a set of ontologies to support semantic browsing. Thanks to Watson, PowerMagpie can dynamically (during browsing) discover ontologies published online that provide the relevant knowledge for the interpretation of the domain terms extracted in the previous step.

3. Relating the text to semantic information. The goal here is to associate the terms of the domain, found in the text of the web pages, with the relevant semantic information. In other term, PowerMagpie has to establish a matching between the entities present in the ontologies (classes, properties and individuals providing semantic descriptions) to the occurrences of terms in the web page.

4. Navigating textual and semantic information together. Once the elements of text have been put in relation with the appropriate semantic information, it is important to provide efficient visualizations and interfaces for interpretation. This requires not only to display semantic information along with the usual textual information provided by the web page, but also to allow for interactions between the user and both forms of information. In a sense, semantic browsing consists in navigating the semantic links relating ontological entities, as well as the relations associating textual elements to semantic interpretations, supporting the user in making the best of the knowledge available online when browsing web documents….

» Read more in the OpenKnowledge Deliverable D8.1 Report.