Improving Multi-hop Question Answering over Knowledge Graphs using Knowledge Base Embeddings

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4498–4507 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 4498 Improving Multi-hop Question Answering over Knowledge Graphs using Knowledge Base Embeddings Apoorv Saxena∗ Aditay Tripathi∗ Partha Talukdar Indian Institute of Science, Bangalore {apoorvsaxena,aditayt,ppt}@iisc.ac.in Abstract Knowledge Graphs (KG) are multi-relational graphs consisting of entities as nodes and rela- tions among them as typed edges. Goal of the Question Answering over KG (KGQA) task is to answer natural language queries posed over the KG. Multi-hop KGQA requires rea- soning over multiple edges of the KG to ar- rive at the right answer. KGs are often in- complete with many missing links, posing ad- ditional challenges for KGQA, especially for multi-hop KGQA. Recent research on multi- hop KGQA has attempted to handle KG spar- sity using relevant external text, which isn’t al- ways readily available. In a separate line of research, KG embedding methods have been proposed to reduce KG sparsity by perform- ing missing link prediction. Such KG em- bedding methods, even though highly relevant, have not been explored for multi-hop KGQA so far. We fill this gap in this paper and pro- pose EmbedKGQA. EmbedKGQA is particu- larly effective in performing multi-hop KGQA over sparse KGs. EmbedKGQA also relaxes the requirement of answer selection from a pre- specified neighborhood, a sub-optimal con- straint enforced by previous multi-hop KGQA methods. Through extensive experiments on multiple benchmark datasets, we demonstrate EmbedKGQA’s effectiveness over other state- of-the-art baselines. 1 Introduction Knowledge Graphs (KG) are multi-relational graphs consisting of millions of entities (e.g., San Jose, California, etc.) and relationships among them (e.g., San Jose-cityInState-California). Ex- amples of a few large KGs include Wikidata (Google, 2013), DBPedia (Lehmann et al., 2015), Yago (Suchanek et al., 2007), and NELL (Mitchell ∗Equal contribution EmbedKGQA’s source code is available at https://github.com/malllabiisc/EmbedKGQA Figure 1: Challenges with Multi-hop QA over Knowl- edge Graphs (KGQA) in sparse and incomplete KGs: Absence of the edge has genre(Gangster No. 1, Crime) in the incomplete KG makes it much harder to answer the input NL question, as the KGQA model potentially needs to reason over a longer path over the KG (marked by bold edges). Existing multi-hop KGQA methods also impose heuristic neighborhood limits (shaded re- gion in the figure), which often makes the true answer (Crime in this example) out of reach. EmbedKGQA, our proposed method, overcomes these limitations by utilizing embeddings of the input KG during multi-hop KGQA. For more details, please refer Figure 2 and Sec- tion 4. et al., 2018). Question Answering over Knowledge Graphs (KGQA) has emerged as an important re- search area over the last few years (Zhang et al., 2018; Sun et al., 2019a). In KGQA systems, given a natural language (NL) question and a KG, the right answer is derived based on analysis of the question in the context of the KG. In multi-hop KGQA, the system needs to perform reasoning over multiple edges of the KG to infer the right answer. …

Content truncated to 500 words