摘要: Multi-document summarization has been an important problem in information retrieval. It aims to distill the most from a set of documents generate compressed summary. Given sentence graph generated where vertices represent sentences and edges indicate that corresponding are similar, extracted summary can be described using idea domination. In this paper, we propose new principled versatile framework for multi-document minimum dominating set. We show four well-known tasks including generic, query-focused, update, comparative modeled as different variations derived proposed framework. Approximation algorithms performing also empirical experiments conducted demonstrate effectiveness our