Multi-Document Scientific Summarization from a Knowledge Graph-Centric View

Proceedings of the 29th International Conference on Computational Linguistics, pages 6222–6233 October 12–17, 2022. 6222 Multi-Document Scientific Summarization from a Knowledge Graph-Centric View Pancheng Wang, Shasha Li†, Kunyuan Pang, Liangliang He, Dong Li, Jintao Tang†, Ting Wang† College of Computer Science and Technology, National University of Defense Technology, Changsha, China {wangpancheng13, shashali, pangkunyuan10, heliangliang19, lidong1,tangjintao, tingwang}@nudt.edu.cn Abstract Multi-Document Scientific Summarization (MDSS) aims to produce coherent and concise summaries for clusters of topic-relevant scientific papers. This task requires precise understanding of paper content and accurate modeling of cross-paper relationships. Knowl- edge graphs convey compact and interpretable structured information for documents, which makes them ideal for content modeling and relationship modeling. In this paper, we present KGSum1, an MDSS model centred on knowledge graphs during both the encoding and decoding process. Specifically, in the encoding process, two graph-based modules are proposed to incorporate knowledge graph information into paper encoding, while in the decoding process, we propose a two-stage decoder by first generating knowledge graph information of summary in the form of descriptive sentences, followed by generating the final summary. Empirical results show that the proposed architecture brings sub- stantial improvements over baselines on the Multi-Xscience dataset. 1 Introduction Nowadays, the exponential increasing publication rate of scientific papers makes it difficult and time-consuming for researchers to keep track of the latest advances. Multi-Document Scientific Summarization (MDSS) is therefore introduced to alleviate this information overload problem by gen- erating succinct and comprehensive summary from clusters of topic-relevant scientific papers (Chen et al., 2021; Shah and Barzilay, 2021). In MDSS, paper content modeling and cross- paper relationship modeling are two main issues. (1) Scientific papers contain complex concepts, technical terms, and abbreviations that convey im- portant information about paper content. However, †Corresponding authors. 1https://github.com/muguruzawang/KGSum some previous works (Wang et al., 2018; Jiang et al., 2019) treat all text units equally, which in- evitably ignore the salient information of some less frequent technical terms and abbreviations. (2) Furthermore, there exist intricate relationships be- tween papers in MDSS, such as sequential, paral- lel, complementary and contradictory (Luu et al., 2021), which play a vital role in guiding the se- lection and organization of different contents. The latest work (Chen et al., 2021) attempt to capture cross-paper relationships via seq2seq model with- out considering any links between fine-grained text units. Failure to take into account explicit relation- ships between papers prevents their model from learning cross-paper relationships effectively. To address the two aforementioned issues, we consider leveraging salient text units, namely enti- ties, and their relations for MDSS. Scientific papers contain multiple domain-specific entities and rela- tions between them. These entities and relations succinctly capture important information about the main content of papers. Knowledge graphs based on these scientific entities and relations can be in- herently used for content modeling of scientific pa- pers. Take Figure 1 as an example. The knowledge graph at the top left illustrates the main content of paper 1, which can be formulated as: Paper 1 uses memory augmented networks method to solve …

Content truncated to 500 words