Knowledge Graph Question Answering Leaderboard: A Community Resource to Prevent a Replication Crisis

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 2998–3007 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 2998 Knowledge Graph Question Answering Leaderboard: A Community Resource to Prevent a Replication Crisis A. Perevalov∗§¶, Xi Yan†¶, L. Kovriguina‡, L. Jiang†, A. Both§ and R. Usbeck† ∗Anhalt University of Applied Sciences, †University of Hamburg, ‡Fraunhofer IAIS, §Leipzig University of Applied Sciences aleksandr.perevalov@hs-anhalt.de, liubov.kovriguina@iais.fraunhofer.de, andreas.both@htwk-leipzig.de, {xi.yan, longquan.jiang, ricardo.usbeck}@uni-hamburg.de ¶These authors contributed equally to this work. Abstract Data-driven systems need to be evaluated to establish trust in the scientific approach and its applicability. In particular, this is true for Knowledge Graph (KG) Question Answering (QA), where complex data structures are made accessible via natural-language interfaces. Evaluating the capabilities of these systems has been a driver for the community for more than ten years while establishing different KGQA benchmark datasets. However, comparing different approaches is cumbersome. The lack of existing and curated leaderboards leads to a missing global view over the research field and could inject mistrust into the results. In particular, the latest and most-used datasets in the KGQA community, LC-QuAD and QALD, miss providing central and up-to-date points of trust. In this paper, we survey and analyze a wide range of evaluation results with significant coverage of 100 publications and 98 systems from the last decade. We provide a new central and open leaderboard for any KGQA benchmark dataset as a focal point for the community - https://kgqa.github.io/leaderboard/. Our analysis highlights existing problems during the evaluation of KGQA systems. Thus, we will point to possible improvements for future evaluations. Keywords: Evaluation Methodology, Knowledge Graph, Question Answering, Leaderboard, Replication Crisis 1. Introduction Question Answering (QA) is a rapidly growing field in research and industry1. QA systems already de- liver their potential into many real-world problems, e.g., (Mutabazi et al., 2021; Both et al., 2021; Diefen- bach et al., 2021). These systems can be divided into two main paradigms (Jurafsky and Martin, 2018): IR- based that works over unstructured data, closely related to Machine Reading Comprehension and Retriever- Reader architecture) and Knowledge-Based (KBQA) which works over structured data, such as relational tables, specific data APIs, knowledge graphs (KGs). In this regard, Question Answering over Knowledge Graphs (KGQA) is of particular interest to this work. Many different benchmarking datasets are used for evaluating KGQA systems. These datasets differ in the underlying knowledge graph (e.g., DBpedia (Auer et al., 2007) or Wikidata (Erxleben et al., 2014)), size or- der of magnitude (Fu et al., 2020), questions complex- ity (Saleem et al., 2017), multilingual support (Chandra et al., 2021), and many more dimensions. In the KGQA research community, several datasets have become a de facto standard for evaluation of such systems, such as the QALD (Usbeck et al., 2018) and LC-QuAD (Dubey et al., 2019) benchmark dataset series. As more and more researchers introduce new evaluation results us- ing these well-known datasets, it becomes more chal- lenging to follow the up-to-date state-of-the-art in the KGQA field. The related …

Content truncated to 500 words