Ernesto William De Luca

| ITC-IRST | EURAC | OVGU-Magdeburg | TU-Berlin | FHP | GEI | All |

Potsdam University
of Applied Sciences

  • Project: Semantic Clustering ( 2013 - Current )

    Since 2013 I cooperate with the Landau Media company (, one of Germany's leading media monitoring and analysis company. Since in times of social media, the number of sources and information is constantly increasing, a review of the relevant information is always time-consuming. In the project "Semantic Clustering for Trend Detection and Sentiment Analysis" we develop a framework to recognize new topics and cluster similar news articles for social media and traditional online media. News articles are processed in order to disambiguate the main concepts and entities contained to simplify the screening of the media response to each topic. The system is trained to automatically recognize if a new article corresponds to a known event, or constitutes a novel event (Topic Detection and Tracking). The articles are grouped into events using a hierarchical agglomerative clustering approach. We also implement a classification approach to categorize news articles into given classes. All the results will be then displayed in an online portal. .
  • Project: SemRes ( 2012 - Current )

    In conservation the documentation is an integral part of the restorers’ work to describe their restored objects. It is important and indispensable to understand the phases of conservation and to allow decision-making for conservation and conservation procedures. The information they contain about the conducted conservation procedures, as well as the methods and the materials used are an immeasurable store of knowledge for this discipline and build the basis of the knowledge used for the conservation of a given object. Documentation is very time consuming; conservators have to categorize photos, give detailed description of the procedures and justify decisions about materials. Every conservator delivers his/her own documentation to the respective monument administrations that archive it in their shelves, without any possibility to access them digitally. In this case, the access to already conducted procedures, as well as the provision of the knowledge about the methods and materials used are very difficult, often almost impossible. At the moment, it is not possible to link different documentations that contain information about similar objects, methods or materials used in the same way. These properties are not recognizable, because of this missing link between them. In the project SemRes (Semantic Documentation for Conservators) I cooperate since 2012 with the River Byte company ( in order to develop a system that can help the conservators in structuring their available documentations accessing them in a semantic way .