Jörg Waitelonis

GERBIL: General Entity Annotator Benchmarking Framework

We present GERBIL, an evaluation framework for semantic entity annotation. The rationale behind our framework is to provide developers, end users and researchers with easy-to-use interfaces that allow for the agile, fine-grained and uniform evaluation of annotation tools on multiple datasets. By these means, we aim to ensure that both tool developers and end users can derive meaningful insights pertaining to the extension, integration and use of annotation applications. In particular, GERBIL provides comparable results to tool developers so as to allow them to easily discover the strengths and weaknesses of their implementations with respect to the state of the art. With the permanent experiment URIs provided by our framework, we ensure the reproducibility and archiving of evaluation results. Moreover, the framework generates data in machine-processable format, allowing for the efficient querying and post-processing of evaluation results. Finally, the tool diagnostics provided by GERBIL allows deriving insights pertaining to the areas in which tools should be further refined, thus allowing developers to create an informed agenda for extensions and end users to detect the right tools for their purposes. GERBIL aims to become a focal point for the state of the art, driving the research agenda of the community by presenting comparable objective evaluation results.

Towards exploratory video search using linked data

Keyword-based search in general is particularly applicable if the searcher really knows what she is looking for and how to find it, i.e. to know the appropriate keywords to obtain the desired results. But in many cases either the objectives of the searcher are intrinsically fuzzy or she is not aware of the appropriate keywords. One way to solve this problem is to navigate and explore the search space along guided routes. In this paper we show, how Linked Open Data can be adopted to facilitate an exploratory semantic search for video data. We present a prototype implementation of exploratory video search and show how traditional keyword-based search can be augmented by the use of Linked Open Data.

WhoKnows? Evaluating linked data heuristics with a quiz that cleans up DBpedia

Purpose – Linking Open Data (LOD) provides a vast amount of well structured semantic information, but many inconsistencies may occur, especially if the data are generated with the help of automated methods. Data cleansing approaches enable detection of inconsistencies and overhauling of affected data sets, but they are difficult to apply automatically. The purpose of this paper is to present WhoKnows?, an online quiz that generates different kinds of questionnaires from DBpedia data sets.Design/methodology/approach – Besides its playfulness, WhoKnows? has been developed for the evaluation of property relevance ranking heuristics on DBpedia data, with the convenient side effect of detecting inconsistencies and doubtful facts.Findings – The original purpose for developing WhoKnows? was to evaluate heuristics to rank LOD properties and thus, obtain a semantic relatedness between entities according to the properties by which they are linked. The presented approach is an efficient method to detect popular prop...

Towards Exploratory Video Search Using Linked Data

Keyword-based search in general is particularly applicable if the searcher really knows what she is looking for and how to find it. But in many cases either the objectives of the searcher are intrinsically fuzzy or she has no idea of the appropriate keywords. One way to solve this problem is to navigate and explore the search space along a guided route. In this paper we show, how Linked Open Data can be adopted to facilitate an exploratory semantic search for video data. We present a prototype implementation of exploratory video search and give first results that show how traditional keyword-based search can be augmented by the use of Linked Open Data.

CONTENTUS--technologies for next generation multimedia libraries

An ever-growing amount of digitized content urges libraries and archives to integrate new media types from a large number of origins such as publishers, record labels and film archives, into their existing collections. This is a challenging task, since the multimedia content itself as well as the associated metadata is inherently heterogeneous—the different sources lead to different data structures, data quality and trustworthiness. This paper presents the contentus approach towards an automated media processing chain for cultural heritage organizations and content holders. Our workflow allows for unattended processing from media ingest to availability thorough our search and retrieval interface. We aim to provide a set of tools for the processing of digitized print media, audio/visual, speech and musical recordings. Media specific functionalities include quality control for digitization of still image and audio/visual media and restoration of the most common quality issues encountered with these media. Furthermore, the contentus tools include modules for content analysis like segmentation of printed, audio and audio/visual media, optical character recognition (OCR), speech-to-text transcription, speaker recognition and the extraction of musical features from audio recordings, all aimed at a textual representation of information inherent within the media assets. Once the information is extracted and transcribed in textual form, media independent processing modules offer extraction and disambiguation of named entities and text classification. All contentus modules are designed to be flexibly recombined within a scalable workflow environment using cloud computing techniques. In the next step analyzed media assets can be retrieved and consumed through a search interface using all available metadata. The search engine combines Semantic Web technologies for representing relations between the media and entities such as persons, locations and organizations with a full-text approach for searching within transcribed information gathered through the preceding processing steps. The contentus unified search interface integrates text, images, audio and audio/visual content. Queries can be narrowed and expanded in an exploratory manner, search results can be refined by disambiguating entities and topics. Further, semantic relationships become not only apparent, but can also be navigated.

Use what you have: Yovisto video search engine takes a semantic turn

The phenomenal increase of online video content confronts the consuming user with an immeasurable amount of data which can only be accessed with sophisticated multimedia search and management technologies. Usual video search engines provide a keyword-based search, where lexical ambiguity of natural language often leads to imprecise and incomplete results. Semantics of keywords and metadata has to be determined to overcome these shortcomings to provide high precision and high recall. In this work, we show how to gradually transform the video search engine Yovisto from a simple keyword-based search engine to a fullyfledged semantic video search engine simply by using the existing search engine infrastructure based on Lucene augmented by simple semantic metadata.

Don't compare Apples to Oranges: Extending GERBIL for a fine grained NEL evaluation

In recent years, named entity linking (NEL) tools were primarily developed as general approaches, whereas today numerous tools are focusing on specific domains such as e.g. the mapping of persons and organizations only, or the annotation of locations or events in microposts. However, the available benchmark datasets used for the evaluation of NEL tools do not reflect this focalizing trend. We have analyzed the evaluation process applied in the NEL benchmarking framework GERBIL [16] and its benchmark datasets. Based on these insights we extend the GERBIL framework to enable a more fine grained evaluation and in deep analysis of the used benchmark datasets according to different emphases. In this paper, we present the implementation of an adaptive filter for arbitrary entities as well as a system to automatically measure benchmark dataset properties, such as the extent of content-related ambiguity and diversity. The implementation as well as a result visualization are integrated in the publicly available GERBIL framework.

Linked Data als Grundlage der semantischen Videosuche mit yovisto

Videodaten sind auf dem besten Wege zur bedeutendsten Informationsquelle im World Wide Web zu werden. Bereits heute werden pro Minute mehr als 100 Stunden Videomaterial (siehe http://www.youtube.com/yt/press/statistics.html, aufgerufen am 01.03.2014.) von den Benutzern auf Videoplattformen wie YouTube eingestellt. Bei dieser gewaltigen Menge an unstrukturierten multimedialen Daten wird auch die gezielte Informationssuche immer schwieriger, da eine inhaltsbasierte Suche mit Hilfe von textbasierten Metadaten realisiert wird, die entweder manuell oder mittels unzuverlassiger automatischer Analyseverfahren gewonnen werden. Hier bietet die semantische Videosuche einen Ausweg, die aufbauend auf einer Vielzahl unterschiedlicher Analyseverfahren versucht, textbasierte Metadaten inhaltlich miteinander in Bezug zu setzen und zielsicher die gewunschten Ergebnisse zu finden. Daruber hinaus ermoglicht es den zu Grunde liegenden Suchraum, d. h. das gesamte Videoarchiv ahnlich dem Stobern in einem gutsortierten Bucherregal zielstrebig zu durchmustern und auf diese Weise hilfreiche neue Informationen zu finden. Die Videosuchmaschine yovisto.com implementiert zahlreiche visuelle Analyseverfahren und kombiniert diese prototypisch in einer explorativen semantischen Suche.

TIB|AV-Portal: Integrating Automatically Generated Video Annotations into the Web of Data

The German National Library of Science and Technology (TIB) aims to promote the use and distribution of its collections. In this context TIB publishes metadata of scientific videos from the TIB|AV Portal as linked open data. Unlike other library metadata the TIB|AV-Portal deploys automated metadata extraction and named entity linking to provide time-based semantic metadata. By publishing this metadata, TIB is offering a new service involving the provision of quarterly updated data in RDF format which can be reused by third parties. In this paper the strategy and the challenges regarding the linked open data service are introduced.

I am a Machine, Let Me Understand Web Media!

The majority of web assets cannot be understood by machines, because of the lack of available explicit and machine readable semantics. By enabling machines to understand the meaning of web media, fully automated discovery, processing, and linking become feasible. Semantic Web technologies offer the possibility to enhance web resources with explicit semantics via linking to ontologies encoded in RDF. We demand to make the content of every web asset explicit for machines with the least possible effort for any content provider. Web servers should deliver RDF descriptions for any web document on request. To achieve this, we propose a framework that enables web content providers to connect to content-wise descriptions of their web assets via simple HTTP content negotiation in connection with on-the-fly automated multimedia analysis services. We demonstrate the feasibility of our approach with a prototype implementation.