Live Blog Corpus for Summarization

Live Blog Corpus for Summarization Avinesh P.V.S., Maxime Peyrard, Christian M. Meyer Research Training Group AIPHES and UKP Lab Computer Science Department, Technische Universit¨at Darmstadt www.aiphes.tu-darmstadt.de, www.ukp.tu-darmstadt.de Abstract Live blogs are an increasingly popular news format to cover breaking news and live events in online journalism. Online news websites around the world are using this medium to give their readers a minute by minute update on an event. Good summaries enhance the value of the live blogs for a reader but are often not available. In this paper, we study a way of collecting corpora for automatic live blog summarization. In an empirical evaluation using well-known state-of-the-art summarization systems, we show that live blogs corpus poses new challenges in the field of summarization. We make our tools publicly available to reconstruct the corpus to encourage the research community and replicate our results. https://github.com/UKPLab/lrec2018-live-blog-corpus Keywords: Live blogs, Summarization Corpus, Corpus Construction, Focused Crawling, Online Journalism 1. Introduction A live blog is a dynamic news article providing a rolling textual coverage of an ongoing event. It is a single arti- cle continuously updated by one or many journalists with timestamped micro-updates typically displayed in chrono- logical order. Live blogs can contain a wide variety of me- dia, including text, video, audio, images, social media snip- pets and links. At the end of the broadcasting, a journalist usually summarizes the main information about the event. For more extended events, journalists may also write in- termediate summaries. Figure 1 and 2 show an example live blog provided by the BBC on “Last day of Supreme Court Brexit Case” and The Guardian on “US elections 2016 campaign”. The timestamped information snippets are on the right, the human-written bullet-point summary is at the top left. Figure 1: BBC.com live blog on “Last day of Supreme Court Brexit Case” In the last decade, live-blogging has become very popular. It is commonly used by major news organizations, such as the BBC, The Guardian or The New York Times. Several different kinds of events are regularly covered by live blogs, including sport games, elections, ceremonies, protests, con- flicts and natural disasters. Thurman and Schapals (2017, p.1) report a journalist’s view that “live blogs have trans- formed the way we think about news, our sourcing, and ev- Figure 2: TheGuardian.com live blog on “US elections 2016 campaign” erything”. Thanks to this new journalistic trend, many live blogs – and their human-written summaries – are available online and new ones are generated every day. In this work, we propose to leverage this data and in- vestigate the task of automatic live blog summarization by crawling a new dataset. Live blog summarization has more direct applications in Journalism than the traditional but rather artificial tasks of a single document and multi- document summarization. Systems capable of summariz- ing live streams of heterogeneous content can be directly beneficial to users and even assist journalists during their daily work. However, this new task also comes with new challenges. Live blogs are a list of …

Content truncated to 500 words