We investigate the problem of reader-aware multi-document summarization(RA-MDS) and introduce a new dataset for this problem. To tackle RA-MDS, weextend a variational auto-encodes (VAEs) based MDS framework by jointlyconsidering news documents and reader comments. To conduct evaluation forsummarization performance, we prepare a new dataset. We describe the methodsfor data collection, aspect annotation, and summary writing as well asscrutinizing by experts. Experimental results show that reader comments canimprove the summarization performance, which also demonstrates the usefulnessof the proposed dataset. The annotated dataset for RA-MDS is available online.
展开▼
