Abstract. We present a hybrid machine learning approach for information extraction from unstructured documents by integrating a learned classifier based on the Maximum Entropy Mod...
: We demonstrate an approach and an accompanying UNIX toolbox for performing wtrious kinds of Knowledge tT,xlractions and Structuring. The goal is to "practically" enhanc...
Antoine Ogonowski, Marie Luce Herviou, Eva Dauphin
Extraction of entities from ad creatives is an important problem that can benefit many computational advertising tasks. Supervised and semi-supervised solutions rely on labeled da...
Since 1995, a few statistical parsing algorithms have demonstrated a breakthrough in parsing accuracy, as measured against the UPenn TREEBANK as a gold standard. In this paper we ...
Scott Miller, Heidi Fox, Lance A. Ramshaw, Ralph M...
Lexical resources such as WordNet and the EDR electronic dictionary (EDR) have been used in several NLP tasks. Probably partly due to the fact that the EDR is not freely available...
Lonneke van der Plas, Vincenzo Pallotta, Martin Ra...