Complex Question Answering on knowledge graphs using machine translation and multi-task learning

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics, pages 3428–3439 April 19 - 23, 2021. ©2021 Association for Computational Linguistics 3428 Complex Question Answering on knowledge graphs using machine translation and multi-task learning Saurabh Srivastava, Mayur Patidar, Sudip Chowdhury, Puneet Agarwal, Indrajit Bhattacharya, Gautam Shroff TCS Research, New Delhi, India { sriv.saurabh, patidar.mayur, sudip.chowdhury2 puneet.a, gautam.shroff}@tcs.com Abstract Question answering (QA) over a knowledge graph (KG) is a task of answering a natu- ral language (NL) query using the informa- tion stored in KG. In a real-world industrial setting, this involves addressing multiple chal- lenges including entity linking, multi-hop rea- soning over KG, etc. Traditional approaches handle these challenges in a modularized se- quential manner where errors in one module lead to the accumulation of errors in down- stream modules. Often these challenges are inter-related and the solutions to them can rein- force each other when handled simultaneously in an end-to-end learning setup. To this end, we propose a multi-task BERT based Neural Machine Translation (NMT) model to address these challenges. Through experimental anal- ysis, we demonstrate the efficacy of our pro- posed approach on one publicly available and one proprietary dataset. 1 Introduction Question answering on knowledge graphs (KGQA) has mainly been attempted on publicly available KGs such as Freebase Bollacker et al. (2008), DB- Pedia Lehmann et al. (2015), Yago Suchanek et al. (2007), etc. There is also a demand for questions answering on proprietary KGs created by large en- terprises. For example, KGQA, on a) a KG that contains information related to retail products, can help the customers choose the right product for their needs, or b) a KG containing document cata- logs (best practices, white papers, research papers) can help a knowledge worker find a specific piece of information, or c) a KG that stores profiles of var- ious companies can be used to do preliminary anal- ysis before giving them a loan, etc. Our motivat- ing use-case comes from an enterprise system (re- ferred to as LOCA) that is expected to answer users’ questions about the R&D division of an enterprise. Figure 1: Example queries from a real-world dataset LOCA. Column 6 (is SP?) represents whether the queries can be answered via shortest path or not?, all the other columns are self-explanatory. Sample questions from LOCA dataset are shown in Figure 1. The schema of the corresponding KG is shown in Figure 2. Answering such questions often requires a traversal of KG along multiple re- lations which may not form a directed chain graph, and may follow more complex topology as shown for question 5, 7 and 8 in Figure 1. It can also be observed that most often words of the natural lan- guage question (NLQ) and corresponding relations have a weak correlation. Most of the proposed ap- proaches on the KGQA task Bollacker et al. (2008) parse the NLQ and convert it into a structured query and then execute the structured query on the KG to retrieve …

Content truncated to 500 words