Concept maps can be used to concisely represent important information andbring structure into large document collections. Therefore, we study a variantof multi-document summarization that produces summaries in the form of conceptmaps. However, suitable evaluation datasets for this task are currentlymissing. To close this gap, we present a newly created corpus of concept mapsthat summarize heterogeneous collections of web documents on educationaltopics. It was created using a novel crowdsourcing approach that allows us toefficiently determine important elements in large document collections. Werelease the corpus along with a baseline system and proposed evaluationprotocol to enable further research on this variant of summarization.
展开▼