Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Pretrained Transformers for Simple Question Answering over Knowledge Graphs

: Lukovnikov, Denis; Fischer, Asja; Lehmann, Jens

Postprint urn:nbn:de:0011-n-6183115 (300 KByte PDF)
MD5 Fingerprint: a89beee17f09fd3b5a72245b3d157276
Erstellt am: 15.12.2020

Ghidini, C.:
The Semantic Web - ISWC 2019. 18th International Semantic Web Conference. Proceedings. Pt.I : Auckland, New Zealand, October 26–30, 2019
Cham: Springer Nature, 2019 (Lecture Notes in Computer Science 11778)
ISBN: 978-3-030-30792-9 (Print)
ISBN: 978-3-030-30793-6 (Online)
International Semantic Web Conference (ISWC) <18, 2019, Auckland>
Konferenzbeitrag, Elektronische Publikation
Fraunhofer IAIS ()

Answering simple questions over knowledge graphs is a well-studied problem in question answering. Previous approaches for this task built on recurrent and convolutional neural network based architectures that use pretrained word embeddings. It was recently shown that finetuning pretrained transformer networks (e.g. BERT) can outperform previous approaches on various natural language processing tasks. In this work, we investigate how well BERT performs on SimpleQuestions and provide an evaluation of both BERT and BiLSTM-based models in limited-data scenarios.