TY - GEN
T1 - Improving Inductive Link Prediction Using Hyper-relational Facts
AU - Ali, Mehdi
AU - Berrendorf, Max
AU - Galkin, Mikhail
AU - Thost, Veronika
AU - Ma, Tengfei
AU - Tresp, Volker
AU - Lehmann, Jens
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - For many years, link prediction on knowledge graphs (KGs) has been a purely transductive task, not allowing for reasoning on unseen entities. Recently, increasing efforts are put into exploring semi- and fully inductive scenarios, enabling inference over unseen and emerging entities. Still, all these approaches only consider triple-based KGs, whereas their richer counterparts, hyper-relational KGs (e.g., Wikidata), have not yet been properly studied. In this work, we classify different inductive settings and study the benefits of employing hyper-relational KGs on a wide range of semi- and fully inductive link prediction tasks powered by recent advancements in graph neural networks. Our experiments on a novel set of benchmarks show that qualifiers over typed edges can lead to performance improvements of 6% of absolute gains (for the Hits@10 metric) compared to triple-only baselines. Our code is available at https://github.com/mali-git/hyper_relational_ilp.
AB - For many years, link prediction on knowledge graphs (KGs) has been a purely transductive task, not allowing for reasoning on unseen entities. Recently, increasing efforts are put into exploring semi- and fully inductive scenarios, enabling inference over unseen and emerging entities. Still, all these approaches only consider triple-based KGs, whereas their richer counterparts, hyper-relational KGs (e.g., Wikidata), have not yet been properly studied. In this work, we classify different inductive settings and study the benefits of employing hyper-relational KGs on a wide range of semi- and fully inductive link prediction tasks powered by recent advancements in graph neural networks. Our experiments on a novel set of benchmarks show that qualifiers over typed edges can lead to performance improvements of 6% of absolute gains (for the Hits@10 metric) compared to triple-only baselines. Our code is available at https://github.com/mali-git/hyper_relational_ilp.
UR - https://www.scopus.com/pages/publications/85116905779
U2 - 10.1007/978-3-030-88361-4_5
DO - 10.1007/978-3-030-88361-4_5
M3 - Conference contribution
AN - SCOPUS:85116905779
SN - 9783030883607
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 74
EP - 92
BT - The Semantic Web – ISWC 2021 - 20th International Semantic Web Conference, ISWC 2021, Proceedings
A2 - Hotho, Andreas
A2 - Blomqvist, Eva
A2 - Dietze, Stefan
A2 - Fokoue, Achille
A2 - Ding, Ying
A2 - Barnaghi, Payam
A2 - Haller, Armin
A2 - Dragoni, Mauro
A2 - Alani, Harith
PB - Springer Science and Business Media Deutschland GmbH
T2 - 20th International Semantic Web Conference, ISWC 2021
Y2 - 24 October 2021 through 28 October 2021
ER -