We present a novel approach to recognizing Textual nt. Structural features are constructed from abstract tree descriptions, which are automatically extracted from syntactic depend...
Graph-based semi-supervised learning (SSL) algorithms have been successfully used to extract class-instance pairs from large unstructured and structured text collections. However,...
While mutual information-based methods have become popular for image registration, the question of what underlying feature to use is rarely discussed. Instead, it is implicitly as...
This work applies boosted wrapper induction (BWI), a machine learning algorithm for information extraction from semi-structured documents, to the problem of named entity recogniti...
Relations between entities in text have been widely researched in the natural language processing and informationextraction communities. The region connecting a pair of entities (...