Collaborative Computing: Networking, Applications and Worksharing. 13th International Conference, CollaborateCom 2017, Edinburgh, UK, December 11–13, 2017, Proceedings

Research Article

Citation Based Collaborative Summarization of Scientific Publications by a New Sentence Similarity Measure

Download
132 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-00916-8_62,
        author={Chengzhe Yuan and Dingding Li and Jia Zhu and Yong Tang and Shahbaz Wasti and Chaobo He and Hai Liu and Ronghua Lin},
        title={Citation Based Collaborative Summarization of Scientific Publications by a New Sentence Similarity Measure},
        proceedings={Collaborative Computing: Networking, Applications and Worksharing. 13th International Conference, CollaborateCom 2017, Edinburgh, UK, December 11--13, 2017, Proceedings},
        proceedings_a={COLLABORATECOM},
        year={2018},
        month={10},
        keywords={Collaborative summarization Sentence similarity measure Citation context},
        doi={10.1007/978-3-030-00916-8_62}
    }
    
  • Chengzhe Yuan
    Dingding Li
    Jia Zhu
    Yong Tang
    Shahbaz Wasti
    Chaobo He
    Hai Liu
    Ronghua Lin
    Year: 2018
    Citation Based Collaborative Summarization of Scientific Publications by a New Sentence Similarity Measure
    COLLABORATECOM
    Springer
    DOI: 10.1007/978-3-030-00916-8_62
Chengzhe Yuan1, Dingding Li1,*, Jia Zhu1, Yong Tang1, Shahbaz Wasti1, Chaobo He1, Hai Liu1, Ronghua Lin1
  • 1: South China Normal University
*Contact email: dingdingli@m.scnu.edu.cn

Abstract

Next-generation network offers unrestricted access for researchers to all kinds of scientific publications, collaborative summarization systems are now being contemplated as a service that can help researchers gain information when they read scientific articles. One way to develop a collaborative summarization system is to measure semantic similarity between sentences to improve its quality. In this paper, we introduce a new sentence similarity measure for summarizing scientific articles with citation context. Our work is based on recent work in document distance metric called the word mover’s distance (WMD). Compared to traditional similarity measures, WMD based sentence similarity measure has better performance by capturing the semantic relation between two sentences. Experiments on 2016 version of ACL Anthology Reference Corpus show that our approach outperforms several other baselines by ROUGE metrics.