A Brief Survey of Multimedia Annotation Localisation on the Web of Linked Data Gary Lefman 1 David Lewis 1 Felix Sasaki 2 1 CNGL the Centre for Global Intelligent Content, Trinity College Dublin, Ireland 2 Language Technology Lab, German Research Centre for Artificial Intelligence (DFKI), Germany
A brief survey feasibility of multilingual annotation on social media using linked data very ^
Why are we looking at this? role of localisation multimedia ontologies linking annotated resources across languages
Rectangle:140,170,50,50 Text:“tree” Author:“Jane Doe” Locale: tree coeden Rectangle:212,182,50,50 Text:“coeden” Author:“John Doe” Locale: Search for:“Baum” Locale: de-DE ? ? “ Baum ” Social Media Platform A Social Media Platform B Platform B Platform A The annotation disconnect
Annotation No Ontology Resource Folksonomy Monolingual “African Lion” “Afrikanskt Lejon” Flickr
Flickr – what did we learn? monolingual platform no hint of an ontology heterogeneous folksonomies simple keyword tagging
Flickr – what did we learn?
monolingual annotations in siloes uncontrolled “machine tags” namespace:predicate=value
Further research linguistically linked folksonomic multimedia annotation across social media platforms
Search for:“Löwe” Locale: de-DE “ Löwe ” LLOD Further research
Better Utilisation of Dublin Core Integrate Media Fragments URI Utilise MediaONT Further research
Questions? Gary Lefman 1 David Lewis 1 Felix Sasaki 2 1 CNGL the Centre for Global Intelligent Content, Trinity College Dublin, Ireland 2 Language Technology Lab, German Research Centre for Artificial Intelligence (DFKI), Germany
Dublin, Ireland 23 rd – 29 th August 2014