Indicative Tweet Generation: An Extractive Summarization Problem?

Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 138–147, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. Indicative Tweet Generation: An Extractive Summarization Problem? Priya Sidhaye School of Computer Science McGill University Montreal, QC, Canada, H3A 0E9 priya.sidhaye@mail.mcgill.ca Jackie Chi Kit Cheung School of Computer Science McGill University Montreal, QC, Canada, H3A 0E9 jcheung@cs.mcgill.ca Abstract Social media such as Twitter have become an important method of communication, with potential opportunities for NLG to fa- cilitate the generation of social media con- tent. We focus on the generation of in- dicative tweets that contain a link to an ex- ternal web page. While it is natural and tempting to view the linked web page as the source text from which the tweet is generated in an extractive summarization setting, it is unclear to what extent ac- tual indicative tweets behave like extrac- tive summaries. We collect a corpus of indicative tweets with their associated ar- ticles and investigate to what extent they can be derived from the articles using ex- tractive methods. We also consider the im- pact of the formality and genre of the ar- ticle. Our results demonstrate the limits of viewing indicative tweet generation as extractive summarization, and point to the need for the development of a methodol- ogy for tweet generation that is sensitive to genre-specific issues. 1 Introduction With the rise in popularity of social media, mes- sage broadcasting sites such as Twitter and other microblogging services have become an important means of communication, with an estimated 500 million tweets being written every day1. In addi- tion to individual users, various organizations and public figures such as newspapers, government officials and entertainers have established them- selves on social media in order to disseminate in- formation or promote their products. While there has been recent progress in the development of Twitter-specific POS taggers, 1https://about.twitter.com/company parsers, and other tweet understanding tools (Owoputi et al., 2013; Kong et al., 2014), there has been little work on methods for generating tweets, despite the utility this would have for users and organizations. In this paper, we study the generation of the par- ticular class of tweets that contain a link to an ex- ternal web page that is composed primarily of text. Given the short length of a tweet, the presence of a URL in the tweet is a strong signal that the tweet is functioning to help Twitter users decide whether to read the full article. This class of tweets, which we call indicative tweets, represents a large sub- set of tweets overall, constituting more than half of the tweets in our data set. Indicative tweets would appear to be the easiest to handle using cur- rent methods in text summarization, because there is a clear source of input from which a tweet could be generated. In effect, the tweet would be acting as an indicative summary of the article it is being linked to, and it would seem that existing meth- ods …

Content truncated to 500 words