Incorporating External Knowledge into Machine Reading for Generative Question Answering

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 2521–2530, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 2521 Incorporating External Knowledge into Machine Reading for Generative Question Answering Bin Bi, Chen Wu, Ming Yan, Wei Wang, Jiangnan Xia, Chenliang Li Alibaba Group {b.bi, wuchen.wc, ym119608, hebian.ww}@alibaba-inc.com {jiangnan.xjn, lcl193798}@alibaba-inc.com Abstract Commonsense and background knowledge is required for a QA model to answer many non- trivial questions. Different from existing work on knowledge-aware QA, we focus on a more challenging task of leveraging external knowl- edge to generate answers in natural language for a given question with context. In this paper, we propose a new neural model, Knowledge-Enriched Answer Gener- ator (KEAG), which is able to compose a natural answer by exploiting and aggregating evidence from all four information sources available: question, passage, vocabulary and knowledge. During the process of answer gen- eration, KEAG adaptively determines when to utilize symbolic knowledge and which fact from the knowledge is useful. This allows the model to exploit external knowledge that is not explicitly stated in the given text, but that is relevant for generating an answer. The empiri- cal study on public benchmark of answer gen- eration demonstrates that KEAG improves an- swer quality over models without knowledge and existing knowledge-aware models, con- firming its effectiveness in leveraging knowl- edge. 1 Introduction Question Answering (QA) has come a long way from answer sentence selection, relational QA to machine reading comprehension. The next- generation QA systems can be envisioned as the ones which can read passages and write long and abstractive answers to questions. Different from extractive question answering, generative QA based on machine reading produces an answer in true natural language which does not have to be a sub-span in the given passage. Most existing models, however, answer ques- tions based on the content of given passages as the only information source. As a result, they may not be able to understand certain passages or to answer certain questions, due to the lack of commonsense and background knowledge, such as the knowl- edge about what concepts are expressed by the words being read (lexical knowledge), and what relations hold between these concepts (relational knowledge). As a simple illustration, given the passage: State officials in Hawaii on Monday said they have once again checked and confirmed that President Barack Obama was born in Hawaii. to answer the question: Was Barack Obama born in the U.S.?, one must know (among other things) that Hawaii is a state in the U.S., which is external knowledge not present in the text corpus. Therefore, a QA model needs to be enriched with external knowledge properly to be able to an- swer many nontrivial questions. Such knowledge can be commonsense knowledge or factual back- ground knowledge about entities and events that is not explicitly expressed but can be found in a knowledge base such as ConceptNet (Speer et al., 2016), Freebase (Pellissier …

Content truncated to 500 words