Towards Comprehensive Description Generation from Factual Attribute-value Tables

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5985–5996 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 5985 Towards Comprehensive Description Generation from Factual Attribute-value Tables Tianyu Liu1, Fuli Luo1, Pengcheng Yang1, Wei Wu1, Baobao Chang1,2 and Zhifang Sui1,2 1MOE Key Lab of Computational Linguistics, School of EECS, Peking University 2Peng Cheng Laboratory, Shenzhen, China {tianyu0421, luofuli, yang pc, wu.wei, chbb, szf}@pku.edu.cn Abstract The comprehensive descriptions for factual attribute-value tables, which should be accu- rate, informative and loyal, can be very helpful for end users to understand the structured data in this form. However previous neural gen- erators might suffer from key attributes miss- ing, less informative and groundless informa- tion problems, which impede the generation of high-quality comprehensive descriptions for tables. To relieve these problems, we first pro- pose force attention (FA) method to encour- age the generator to pay more attention to the uncovered attributes to avoid potential key at- tributes missing. Furthermore, we propose re- inforcement learning for information richness to generate more informative as well as more loyal descriptions for tables. In our experi- ments, we utilize the widely used WIKIBIO dataset as a benchmark. Additionally we cre- ate WB-filter based on WIKIBIO to test our model in the simulated user-oriented sce- narios, in which the generated descriptions should accord with particular user interests. Experimental results show that our model out- performs the state-of-the-art baselines on both automatic and human evaluation. 1 Introduction Generating descriptions for the factual attribute- value tables has attracted widely interests among NLP researchers especially in a neural end-to-end fashion (e.g. Lebret et al. (2016); Liu et al. (2018); Sha et al. (2018); Bao et al. (2018); Puduppully et al. (2018); Li and Wan (2018); Nema et al. (2018)) as shown in Fig 1a. For broader potential applications in this field, we also simulate user- oriented generation, whose goal is to provide com- prehensive generation for the selected attributes according to particular user interests like Fig 1b. However, we find that previous models might miss key information and generate less informa- Attribute Value Birthplace Utah, America Position forward (soccer player) Comprehensive: A Utah soccer player who plays as forward Missing Key Attri.: A soccer player who plays as forward Groundless info: A Utah forward in the national team Less Informative: An American forward Table 1: An example for comprehensive generation. Suppose we only have two attribute-value tuples, the underlined content is groundless information not men- tioned in source tables. tive and groundless content in its generated de- scriptions towards source tables. For example, in Table 1, the ‘missing key attribute’ case doesn’t mention where the player comes from (birthplace) while the ‘less informative’ one chooses American rather than Utah. The case with groundless infor- mation contains ‘in the national team’ which is not mentioned in the source attributes. Although the ‘key points missing’ problem exists in many text-to-text and data-to-text datasets, for large- scale structured tables with vast heterogeneous at- tributes such …

Content truncated to 500 words