Automatic Creation and Translation of Concept Maps for Computer Science-related Theses and Dissertations

Concept maps are often used as tools to enhance student learning on new topics. We hypothesize that they also can be used as summarization tools for large documents. We hypothesize further that, because concept maps usually have short text sections (words and short phrases), it should be easier to translate concept maps than parts of documents (such as abstracts), since the text need not be natural-sounding sentences. We have selected electronic theses and dissertations (ETDs) as our sample collection, since we have large numbers of these in both English and Spanish. We have adapted a tool that automatically produces semantic maps from text, called Relex, to produce concept maps of ETDs for computing documents, using an ontology of computer science. We also have developed code to automatically find translations for English phrases by using a Spanish ETD collection as a corpus.