A Temporally Sensitive Submodularity Framework for Timeline Summarization

Proceedings of the 22nd Conference on Computational Natural Language Learning (CoNLL 2018), pages 230–240 Brussels, Belgium, October 31 - November 1, 2018. c⃝2018 Association for Computational Linguistics 230 A Temporally Sensitive Submodularity Framework for Timeline Summarization Sebastian Martschat∗ Knowledge Architecture & Innovation BASF SE 67056 Ludwigshafen am Rhein, Germany sebastian.martschat@basf.com Katja Markert Institute of Computational Linguistics Heidelberg University 69120 Heidelberg, Germany markert@cl.uni-heidelberg.de Abstract Timeline summarization (TLS) creates an overview of long-running events via dated daily summaries for the most important dates. TLS differs from standard multi-document summarization (MDS) in the importance of date selection, interdependencies between summaries of different dates and by having very short summaries compared to the num- ber of corpus documents. However, we show that MDS optimization models using submod- ular functions can be adapted to yield well- performing TLS models by designing objec- tive functions and constraints that model the temporal dimension inherent in TLS. Impor- tantly, these adaptations retain the elegance and advantages of the original MDS mod- els (clear separation of features and inference, performance guarantees and scalability, little need for supervision) that current TLS-specific models lack. An open-source implementation of the framework and all models described in this paper is available online.1 1 Introduction There is an abundance of reports on events, crises and disasters. Timelines (see Table 1) summarize and date these reports in an ordered overview. Au- tomatic Timeline Summarization (TLS) constructs such timelines from corpora that contain articles about the corresponding event. In contrast to standard multi-document sum- marization (MDS), in TLS we need to explic- itly model the temporal dimension of the task, specifically we need to select the most important dates for a long-running event and summarize each of these dates. In addition, TLS deals with a much larger number of documents to summarize, ∗Work conducted while the author was a researcher at the Institute of Computational Linguistics, Heidelberg Uni- versity. 1http://smartschat.de/software 2011-03-16 Security forces break up a gathering in Marjeh Square in Damascus of 150 protesters holding pictures of imprisoned relatives. Witnesses say 30 people are arrested. 2011-03-24 President Bashar al-Assad orders the formation of a com- mittee to study how to raise living standards and lift the law covering emergency rule, in place for 48 years. 2011-03-29 Government resigns. Table 1: Excerpt from a Syrian War Reuters timeline. enhancing scalability and redundancy problems. These differences have significant consequences for constraints, objectives, compression rates and scalability (see Section 2.2). Due to these differences, most work on TLS has been separate from the MDS community.2 In- stead, approaches to TLS start from scratch, opti- mizing task-specific heuristic criteria (Chieu and Lee, 2004; Yan et al., 2011b; Wang et al., 2016, inter alia), often with manually determined param- eters (Chieu and Lee, 2004; Yan et al., 2011b) or needing supervision (Wang et al., 2016). As fea- tures and architectures are rarely reused or indeed separated from each other, it is difficult to assess reported improvements. Moreover, none of these approaches give performance guarantees for the task, which are possible …

Content truncated to 500 words