Robust Information Extraction with Perceptrons
Mihai Surdeanu, Massimiliano Ciaramita · 2007
We present a system for the extraction of entity and relation mentions. Our work focused on robustness and simplic-ity: all system components are modeled using variants of the Perceptron algo-rithm (Rosemblatt, 1858) and only partial syntactic information is used for feature extraction. Our approach has two novel ideas. First, we define a new large-margin Perceptron algorithm tailored for class-unbalanced data which dynamically ad-justs its margins, according to the gener-alization performance of the model. Sec-ond, we propose a novel architecture that lets classification ambiguities flow through the system and solves them only at the end. The system achieves compet-itive accuracy on the ACE English EMD and RMD tasks.