Multi-Document Biography Summarization

Multi-document Biography Summarization Liang Zhou, Miruna Ticrea, Eduard Hovy University of Southern California Information Sciences Institute 4676 Admiralty Way Marina del Rey, CA 90292-6695 {liangz, miruna, hovy} @isi.edu Abstract In this paper we describe a biography summarization system using sentence classification and ideas from information retrieval. Although the individual techniques are not new, assembling and applying them to generate multi-document biographies is new. Our system was evaluated in DUC2004. It is among the top performers in task 5–short summaries focused by person questions. 1 Introduction Automatic text summarization is one form of information management. It is described as selecting a subset of sentences from a document that is in size a small percentage of the original and yet is just as informative. Summaries can serve as surrogates of the full texts in the context of Information Retrieval (IR). Summaries are created from two types of text sources, a single document or a set of documents. Multi-document summarization (MDS) is a natural and more elaborative extension of single-document summarization, and poses additional difficulties on algorithm design. Various kinds of summaries fall into two broad categories: generic summaries are the direct derivatives of the source texts; special- interest summaries are generated in response to queries or topic-oriented questions. One important application of special-interest MDS systems is creating biographies to answer questions like “who is Kofi Annan?”. This task would be tedious for humans to perform in situations where the information related to the person is deeply and sparsely buried in large quantity of news texts that are not obviously related. This paper describes a MDS biography system that responds to the “who is” questions by identifying information about the person-in- question using IR and classification techniques, and creates multi-document biographical summaries. The overall system design is shown in Figure 1. To determine what and how sentences are selected and ranked, a simple IR method and experimental classification methods both contributed. The set of top-scoring sentences, after redundancy removal, is the resulting biography. As yet, the system contains no inter-sentence ‘smoothing’ stage. In this paper, work in related areas is discussed in Section 2; a description of our biography corpus used for training and testing the classification component is in Section 3; Section 4 explains the need and the process of classifying sentences according to their biographical state; the application of the classification method in biography extraction/summarization is described in Section 5; an accompanying evaluation on the quality of the biography summaries is shown in Section 6; and future work is outlined in Section 7. 2 Recent Developments Two trends have dominated automatic summarization research (Mani, 2001). One is the work focusing on generating summaries by extraction, which is finding a subset of the document that is indicative of its contents (Kupiec et al., 1995) using “shallow” linguistic analysis and statistics. The other influence is the exploration of Figure 1. Overall design of the biography summarization system. “deeper” knowledge-based methods for condensing information. Knight and Marcu (2000) equate summarization with compression at sentence …

Content truncated to 500 words