Anecdote Recognition and Recommendation

Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, pages 2592–2602, Osaka, Japan, December 11-17 2016. Anecdote Recognition and Recommendation Wei Song†, Ruiji Fu‡, Lizhen Liu†, Hanshi Wang†, Ting Liu§ †Information Engineering, Capital Normal University, Beijing ‡Iflytek Research Beijing, Beijing §Harbin Institute of Technology, Harbin {wsong, lzliu, hswang}@cnu.edu.cn, rjfu@iflytek.com, tliu@ir.hit.edu.cn Abstract We introduce a novel task Anecdote Recognition and Recommendation. An anecdote is a story with a point revealing account of an individual person. Recommending proper anecdotes can be used as evidence to support argumentative writing or as a clue for further reading. We represent an anecdote as a structured tuple — < person, story, implication >. Anecdote recognition runs on archived argumentative essays. We extract narratives containing events of a person as the anecdote story. More importantly, we uncover the anecdote implication, which reveals the meaning and topic of an anecdote. Our approach depends on discourse role identification. Discourse roles such as thesis, main ideas and support help us locate stories and their implications in essays. The experiments show that informative and interpretable anecdotes can be recognized. These anecdotes are used for anecdote recommendation. The anecdote recommender can recommend proper anecdotes in response to given topics. The anecdote implications contribute most for bridging user interested topics and relevant anecdotes. 1 Introduction Building technical tools to assist learning and writing is of great significance and challenging. While a number of tools have been developed for giving feedback on spelling (Bangert-Drowns, 1993), grammar patterns (Yen et al., 2015) and organization (Burstein et al., 2003b), little exists to provide support during planning and composition process. During the process, an automated system, that can effectively collect topic oriented evidence and reading materials, will greatly reduce the cognitive load. This paper introduces the anecdote recognition and recommendation task. An anecdote is a story with a point revealing account of an individual person or an incident. We aim to recognize anecdotes from texts and recommend anecdotes according to given topics. The recommended anecdotes can be used as evidence to support argumentative writing. The argumentative essay is a genre of writing that requires the writer to investigate a topic and establish a position in a concise manner. In addition to create good claims, the quality of argument is greatly affected by the effectiveness of evidence. Finding relevant evidence is not easy, since it heavily depends on long-term accumulation of materials (from reading and observation) and the ability to retrieve and figure out right ones from memory. This process brings in great challenges for both novice and more sophisticated writers. Anecdotal evidence is one of the most commonly used evidence types (Hornikx, 2005). For a given claim, a system that can provide relevant anecdotes would help writers find good evidence and potentially improve the organization and the quality of essays. Recommending anecdotes can be the first step to recommend all types of evidence. Recognizing anecdotes is related to previous work that extracts story-like elements from text. For example, Chambers and Jurafsky (2008) propose …

Content truncated to 500 words