Data-to-text Generation with Entity Modeling

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2023–2035 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 2023 Data-to-text Generation with Entity Modeling Ratish Puduppully and Li Dong and Mirella Lapata Institute for Language, Cognition and Computation School of Informatics, University of Edinburgh 10 Crichton Street, Edinburgh EH8 9AB r.puduppully@sms.ed.ac.uk li.dong@ed.ac.uk mlap@inf.ed.ac.uk Abstract Recent approaches to data-to-text generation have shown great promise thanks to the use of large-scale datasets and the application of neural network architectures which are trained end-to-end. These models rely on represen- tation learning to select content appropriately, structure it coherently, and verbalize it gram- matically, treating entities as nothing more than vocabulary tokens. In this work we pro- pose an entity-centric neural architecture for data-to-text generation. Our model creates entity-specific representations which are dy- namically updated. Text is generated con- ditioned on the data input and entity mem- ory representations using hierarchical atten- tion at each time step. We present experi- ments on the ROTOWIRE benchmark and a (five times larger) new dataset on the baseball domain which we create. Our results show that the proposed model outperforms competitive baselines in automatic and human evaluation.1 1 Introduction Data-to-text generation is the task of generating textual output from non-linguistic input (Reiter and Dale, 1997; Gatt and Krahmer, 2018). The in- put may take on several guises including tables of records, simulations of physical systems, spread- sheets, and so on. As an example, Figure 1 shows (in a table format) the scoring summary of a major league baseball (MLB) game, a play-by-play sum- mary with details of the most important events in the game recorded chronologically (i.e., in which play), and a human-written summary. Modern approaches to data-to-text generation have shown great promise (Lebret et al., 2016; Mei et al., 2016; Perez-Beltrachini and Lapata, 2018; Puduppully et al., 2019; Wiseman et al., 1Our code and dataset can be found at https:// github.com/ratishsp/data2text-entity-py. 2017) thanks to the use of large-scale datasets and neural network models which are trained end-to- end based on the very successful encoder-decoder architecture (Bahdanau et al., 2015). In con- trast to traditional methods which typically imple- ment pipeline-style architectures (Reiter and Dale, 2000) with modules devoted to individual genera- tion components (e.g., content selection or lexical choice), neural models have no special-purpose mechanisms for ensuring how to best generate a text. They simply rely on representation learning to select content appropriately, structure it coher- ently, and verbalize it grammatically. In this paper we are interested in the genera- tion of descriptive texts such as the game summary shown in Figure 1. Descriptive texts are often characterized as “entity coherent” which means that their coherence is based on the way entities (also known as domain objects or concepts) are introduced and discussed in the discourse (Kara- manis et al., 2004). Without knowing anything about baseball or how game summaries are typi- cally written, a glance at the text in Figure 1 re- veals that it is …

Content truncated to 500 words