Ensuring Readability and Data-fidelity using Head-modifier Templates in Deep Type Description Generation

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2036–2046 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 2036 Ensuring Readability and Data-fidelity using Head-modifier Templates in Deep Type Description Generation Jiangjie Chen†, Ao Wang†, Haiyun Jiang†, Suo Feng†, Chenguang Li†, Yanghua Xiao†‡∗ †Shanghai Key Laboratory of Data Science, School of Computer Science, Fudan University, China ‡Shanghai Institute of Intelligent Electronics & Systems, Shanghai, China {jiangjiechen14, awang15, jianghy16, fengs17, cgli17, shawyh}@fudan.edu.cn Abstract A type description is a succinct noun com- pound which helps human and machines to quickly grasp the informative and distinctive information of an entity. Entities in most knowledge graphs (KGs) still lack such de- scriptions, thus calling for automatic meth- ods to supplement such information. How- ever, existing generative methods either over- look the grammatical structure or make fac- tual mistakes in generated texts. To solve these problems, we propose a head-modifier template-based method to ensure the readabil- ity and data fidelity of generated type descrip- tions. We also propose a new dataset and two automatic metrics for this task. Experiments show that our method improves substantially compared with baselines and achieves state- of-the-art performance on both datasets. 1 Introduction Large-scale open domain KGs such as DBpe- dia (Auer et al., 2007), Wikidata (Vrandeˇci´c and Kr¨otzsch, 2014) and CN-DBpedia (Xu et al., 2017) are increasingly drawing the attention from both academia and industries, and have been suc- cessfully used in many applications that require background knowledge to understand texts. In KGs, a type description (Bhowmik and de Melo, 2018) is a kind of description which re- flects the rich information of an entity with lit- tle cognitive efforts. A type description must be informative, distinctive and succinct to help human quickly grasp the essence of an unfa- miliar entity. Compared to other kinds of data in a KG, types in entity-typing task (Shimaoka et al., 2016; Ren et al., 2016) are too general and not informative enough (e.g., when asked about “what is rue Cazotte?”, street in Paris, France is obviously more informative and distinctive than a type location.), and the fixed ∗Corresponding author: Yanghua Xiao. ��� �������� ����� ��� �������� �� ������ ���� ����� ����� ������� ������� ��� ������� ������ ���� ������� �� ��� �������������� ����������� ������ ����� ��� ������ ������ ���� ��� �������������� $hed$ in $mod$ , $mod$ street in Paris , France Infobox of rue Cazotte Stage 2 Stage 1 Generate template Generate type description Figure 1: An example of the two-stage generation of our head-modifier template-based method. $hed$ and $mod$ are the placeholder for head and modifier com- ponents in the template. type set is too inflexible to expand; while infobox and abstract are too long with too much informa- tion, which increases cognitive burden. Type descriptions are useful for a wide range of applications, including question answering (e.g. what is rue Cazotte?), named entity disambigua- tion (e.g. Apple (fruit of the apple tree) vs Apple (American technology company)), taxonomy en- richment, etc. However, …

Content truncated to 500 words