Semi-supervised Abstraction-Augmented String Kernel for Multi-level Bio-Relation Extraction

ervised Abstraction-Augmented String Kernel for Multi-Level Bio-Relation Extraction Pavel Kuksa1 , Yanjun Qi2 , Bing Bai2 , Ronan Collobert2 , Jason Weston3 , Vladimir Pavlovic1 , and Xia Ning4 1 Department of Computer Science, Rutgers University, USA 2 NEC Labs America, Princeton, USA 3 Google Research, New York City, USA 4 Computer Science Department, University of Minnesota, USA Bio-relation extraction (bRE), an important goal in bio-text mining, involves subtasks identifying relationships between bio-entities in text at multiple levels, e.g., at the article, sentence or relation level. A key limitation of current bRE systems is that they are restricted by the availability of annotated corpora. In this work we introduce a semisupervised approach that can tackle multi-level bRE via string comparisons with mismatches in the string kernel framework. Our string kernel implements an abstraction step, which groups similar words to generabstract entities, which can be learnt with unlabeled...

• Type: Journal
• Year: 2010
• Where: PKDD
• Authors: Pavel P. Kuksa, Yanjun Qi, Bing Bai, Ronan Collobert, Jason Weston, Vladimir Pavlovic, Xia Ning
• Keywords: Abstraction-augmented String, Bio-relation Extraction, Data Mining, PKDD 2010, String Kernel