Structured Machine Learning for Biomedical Knowledge Extraction.

The biomedical literature is vast and increasing, and scientists are only able to read a diminishing proportion of it. Only a small fraction of the biomedical knowledge has been extracted and placed in databases. Scientists therefore urgently need better tools to interact with the literature. In addition, while computers have the ability to analyze the literature they are still very limited with regard to knowledge extraction from text. I therefore propose to apply my structured machine learning system to biomedical knowledge extraction. The proposed learning model is able to exploit the available expert knowledge and ontologies. This model is based on a previous model proposed in my PhD thesis for extraction of spatial information from text. This model can deal with the complexity of structured learning from relational data by decomposing a large problem to smaller ones and exploiting efficient optimization techniques for inference. The specific tasks we work on in the Biomedical domain are extraction of the Bacteria and their Habitats and also another task of extraction of biological Pathway information from text.

Keywords:Natural language processing, Natural language meaning representation, Structured machine learning, Biomedical information extraction, Biomedical knowledge extraction Inferenc, Ontology population

Disciplines:Law