RelationListwise for Query-Focused Multi-Document Summarization

Proceedings of COLING 2012: Technical Papers, pages 2961–2976, COLING 2012, Mumbai, December 2012. RelationListwise for query-focused multi-document summarization Wenpeng Y in Li f u Huang Yulong Pei Lian′en Huang The Shenzhen Key Lab for Cloud Computing Technology & Applications (SPCCTA) Shenzhen Graduate School Peking University, Shenzhen 518055, P.R. China {mr.yinwenpeng,warrior.fu,paul.yulong.pei}@gmail.com, hle@net.pku.edu.cn ABSTRACT Most existing learning to rank based summarization methods only used content relevance of sentences with respect to queries to rank or estimate sentences, while neglecting sentence relationships. In our work, we propose a novel model, RelationListwise, by integrating relation information among all the estimated sentences into listMLE-Top K, a basic listwise learning to rank model, to improve the quality of top-ranked sentences. In addition, we present some unique sentence features as well as a novel measure of sentence semantic relation, aiming to enhance the performance of training model. Experimental results on DUC2005-2007 standard summarization data sets demonstrate the effectiveness of our proposed method. KEYWORDS: Query-focused multi-document summarization, Listwise, Sentence relation. 2961 1 Introduction In this paper, we focus on the task of producing extraction-based query-focused multi- document summaries given a collection of documents, which is usually considered as a sen- tence ranking problem. Typically, ranking methods calculate the combinational effects of var- ious features which are designed to identify the different aspects of sentences and/or their relevance to queries. Yet so far not much attention has been paid to it. Most commonly, the features are simply combined by a linear function in which the weights are assigned manu- ally or tuned experimentally. In the past, machine learning approaches have been successfully applied in extractive summarization (Ouyang et al., 2007; Shen and Li, 2011), and a new re- search branch named “learning to rank" has emerged. Its objective is to explore how the optimal weights can be obtained automatically by developing learning strategies. However, previous work mainly considered the content relevance of sentences with respect to certain query while ignoring the relationships among sentences. In this paper, we try to study how to use sentence relatedness to improve the performance of a ranking model. Further, we notice that many learning to rank algorithms have been proposed in recent literature, and these algo- rithms can be categorized into three types: pointwise, pairwise, and listwise approaches. The pointwise and pairwise approaches transform ranking problem into regression or classification on single object and object pairs respectively, while neglecting the fact that ranking is a predic- tion task on a list of objects. In listwise approach, object lists instead of object pairs are used as instances in learning, and the major task is how to construct a listwise loss function, rep- resenting the difference between the ranking list output by a ranking model and the ranking list given as ground truth. Experimental results showed that listwise approach usually outper- forms pointwise and pariwise approaches (Cao et al., 2007; Qin et al., 2008). Accordingly, we mainly concentrate on developing listwise learning to rank in our summarization task. More exactly, taking into account the specific …

Content truncated to 500 words