Towards Interpretable Reasoning over Paragraph Effects in Situation

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 6745–6758, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 6745 Towards Interpretable Reasoning over Paragraph Effects in Situation Mucheng Ren1∗† Xiubo Geng2†, Tao Qin3, Heyan Huang1 and Daxin Jiang2‡ 1School of Computer Science and Technology, Beijing Institute of Technology, Beijing, China 2STCA NLP Group, Microsoft 3Microsoft Research Asia {renm,hhy63}@bit.edu.cn {xigeng,taoqin,djiang}@microsoft.com Abstract We focus on the task of reasoning over para- graph effects in situation, which requires a model to understand the cause and effect de- scribed in a background paragraph, and apply the knowledge to a novel situation. Existing works ignore the complicated reasoning pro- cess and solve it with a one-step “black box” model. Inspired by human cognitive processes, in this paper we propose a sequential approach for this task which explicitly models each step of the reasoning process with neural network modules. In particular, five reasoning mod- ules are designed and learned in an end-to-end manner, which leads to a more interpretable model. Experimental results on the ROPES dataset demonstrate the effectiveness and ex- plainability of our proposed approach. 1 Introduction As a long-standing fundamental task of natural lan- guage processing, machine reading comprehension (MRC) has attracted remarkable attention recently and different MRC datasets have been studied (Ra- jpurkar et al., 2018; Dua et al., 2019b; Choi et al., 2018; Yang et al., 2018), among which reason- ing over paragraph effects in situation (ROPES for short) is a very challenging scenario that needs to understand knowledge from a background para- graph and apply it to answer questions in a novel situation. Table 1 shows an example of the ROPES dataset (Lin et al., 2019), where the background passage states that developmental difficulties could usually be treated by using iodized salt, the situ- ation passage describes two villages using differ- ent salt, and questions about which village having more/less people experiencing developmental diffi- culties need to be answered. ∗Work done during internship at STCA NLP Group, Mi- crosoft. †Equal Contribution ‡Corresponding author Background Before iodized salt was developed, some people experienced a number of developmental diffi- culties, including problems with thyroid gland function and mental retardation. In the 1920s, we learned that these conditions could usually be treated easily with the addition of iodide anion to the diet. One easy way to increase iodide intake was to add the anion to table salt. Situation People from two villages ate lots of salt. People from Salt village used regular salt, while people from Sand village people used iodized salt in their diets, after talking to specialists. Q&A Q: Which village had more people experience developmental difficulties? A: Salt Q: Which village had less people experience developmental difficulties? A: Sand Table 1: An example from the ROPES dataset. Effect property tokens are highlighted in blue, cause property tokens in orange, and world tokens in green. Almost all existing works (Lin et al., 2019; Khashabi et al., 2020; Dua et al., 2019a; Gardner et al., 2020) for this task …

Content truncated to 500 words