F1 is Not Enough! Models and Evaluation Towards User-Centered Explainable Question Answering

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 7076–7095, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 7076 F1 is Not Enough! Models and Evaluation Towards User-Centered Explainable Question Answering Hendrik Schuff1,2 Heike Adel1 Ngoc Thang Vu2 1Bosch Center for Artificial Intelligence, Renningen, Germany 2Institut f¨ur Maschinelle Sprachverarbeitung, University of Stuttgart {Hendrik.Schuff,Heike.Adel}@de.bosch.com Thang.Vu@ims.uni-stuttgart.de Abstract Explainable question answering systems pre- dict an answer together with an explanation showing why the answer has been selected. The goal is to enable users to assess the cor- rectness of the system and understand its rea- soning process. However, we show that cur- rent models and evaluation settings have short- comings regarding the coupling of answer and explanation which might cause serious issues in user experience. As a remedy, we propose a hierarchical model and a new regularization term to strengthen the answer-explanation cou- pling as well as two evaluation scores to quan- tify the coupling. We conduct experiments on the HOTPOTQA benchmark data set and per- form a user study. The user study shows that our models increase the ability of the users to judge the correctness of the system and that scores like F1 are not enough to estimate the usefulness of a model in a practical set- ting with human users. Our scores are bet- ter aligned with user experience, making them promising candidates for model selection. 1 Introduction Understanding the decisions of deep learning mod- els is of utmost importance, especially when they are deployed in critical domains, such as medicine or finance (Ribeiro et al., 2016). In natural lan- guage processing (NLP), a variety of tasks have been addressed regarding explainability of neu- ral networks, such as textual entailment (Camburu et al., 2018), sentiment classification (Clos et al., 2017), machine translation (Stahlberg et al., 2018) and question answering (Yang et al., 2018). In this paper, we address question answering (QA) due to its proximity to users in real-life settings, for instance, in the context of personal assistants. Explainable question answering (XQA) is the task of (i) answering a question and (ii) providing an explanation that enables the user to understand * The Kalahari Desert is a large semi-arid sandy savanna in Southern Africa extending for 900000 km2 , covering much of Botswana, parts of Namibia and regions of South Africa. What is the area of the desert that Ghanzi is in the middle of? 900000 km2, because: * Ghanzi is a town in the middle of the Kalahari Desert the western part of the Republic of Botswana in southern Africa. Ghanzi's area is 117,910 km². Figure 1: Example output of a representative XQA sys- tem (Yang et al., 2018) that would receive an answer-F1 of 1 and an explanation-F1 of 0.5 although the explana- tion provides no value to the user since the actual an- swer evidence (shown in cloud) is not included in the explanation (asterisks mark ground truth explanation). why the answer was selected, e.g., by pointing to the facts that are needed …

Content truncated to 500 words