Semantics and Media
We partook in the organization of the Semantics and Media Workshop at the University of Mainz, where Axel is a Research Fellow. The aim of the workshop was to bring practitioners and research all around media and semantics together and to discuss the application of semantic technologies to achieve media representation, retrieval and convergence. The [...]
FOX Version 0.1
We are thrilled to announce the first version of the Federated knOwledge eXtraction (FOX) framework. FOX integrates and merges the results of frameworks for Named Entity Recognition, Keyword/Keyphrase Extraction and Relation Extraction by using machine learning techniques. By these means, FOX can generate RDF out of natural language with improved accuracy. FOX has been shown [...]
General Overview
Several services and frameworks have been developed to consume natural language and generate semi- to structured data. These tools are based on very diverse algorithms with a variety of strengths and weaknesses. FOX uses this algorithmic diversity to extract RDF of high accuracy out of natural language. It combines tools for natural language processing and aggregates their results by using efficient ensemble learning algorithms.
Evaluation
FOX has been evaluated on several corpora including general corpora such as news and website data sets and on domain-specific corpora from the area of tourism and renewable energy provided within the SCMS project. In all evaluations, FOX outperformed state-of-the-art tools such as the Stanford NER tagger and the Illinois tagger significantly. A full description of the evaluation of FOX on general corpora can be found 
here. THe data used to evaluate FOX's named entity disambiguation approaches can be found 
here.
Contact
 | Dr. Axel-C. Ngonga Ngomo Johannisgasse 26, Zimmer 5-22 04103 Leipzig
| ||||||
E-Mail | Norman Heino Johannisgasse 26, Zimmer 5-08 04103 Leipzig
| ||||||
Information
Last Modification: 
2011-11-16 10:32:21 by Axel Ngonga