Neural Storyline Extraction Model for Storyline Generation from News Articles

Proceedings of NAACL-HLT 2018, pages 1727–1736 New Orleans, Louisiana, June 1 - 6, 2018. c⃝2018 Association for Computational Linguistics Neural Storyline Extraction Model for Storyline Generation from News Articles Deyu Zhou† Linsen Guo† Yulan He§ † School of Computer Science and Engineering, Key Laboratory of Computer Network and Information Integration, Ministry of Education, Southeast University, China § School of Engineering and Applied Science, Aston University, UK {d.zhou,guolinsen}@seu.edu.cn, y.he@cantab.net Abstract Storyline generation aims to extract events de- scribed on news articles under a certain top- ic and reveal how those events evolve over time. Most existing approaches first train su- pervised models to extract events from news articles published in different time periods and then link relevant events into coherent stories. They are domain dependent and cannot deal with unseen event types. To tackle this prob- lem, approaches based on probabilistic graphic models jointly model the generations of events and storylines without annotated data. How- ever, the parameter inference procedure is too complex and models often require long time to converge. In this paper, we propose a nov- el neural network based approach to extrac- t structured representations and evolution pat- terns of storylines without using annotated da- ta. In this model, title and main body of a news article are assumed to share the similar storyline distribution. Moreover, similar doc- uments described in neighboring time periods are assumed to share similar storyline distribu- tions. Based on these assumptions, structured representations and evolution patterns of sto- rylines can be extracted. The proposed mod- el has been evaluated on three news corpora and the experimental results show that it out- performs state-of-the-art approaches accuracy and efficiency. 1 Introduction With the development of the internet, massive in- formation about current events is generated and propagated continuously on online news media sites. It is difficult for the public to digest such large volumes of information effectively. Story- line generation, aiming at summarizing the devel- opment of certain related events, has been inten- sively studied recently (Diao and Jiang, 2014). In general, storyline can be considered as an event cluster where event-related news articles are ordered and clustered depending on both con- tent and temporal similarity. Different ways of calculating content and temporal similarity can be used to cluster related events (Yan et al., 2011; Huang and Huang, 2013). Bayesian non- parametric models could also be used to tackle this problem by describing the storyline gener- ating process using probabilistic graphical mod- els (Li and Cardie, 2014; Diao and Jiang, 2014). Nevertheless, most existing approaches extrac- t events independently and link relevant events in a post-processing step. More recently, Zhou et al. (2016) proposed a non-parametric genera- tive model to extract storylines which is combined with Chinese Restaurant Processes (CRPs) to de- termine the number of storylines automatically. However, the parameter inference procedure is too complex and the model requires long time to con- verge. This makes it impractical to be deployed in real-world applications. Recently, deep learning techniques have been successfully applied to various …

Content truncated to 500 words