X-WikiRE: A Large, Multilingual Resource for Relation Extraction as Machine Comprehension
This addresses the bias in knowledge bases towards English by leveraging multilingual Wikipedia data for relation extraction, though it is incremental as it builds on existing machine comprehension approaches.
The authors tackled the problem of multilingual relation extraction by introducing X-WikiRE, a large dataset that frames it as machine comprehension, enabling cross-lingual model transfer and improving zero-shot relation extraction to populate low-resourced knowledge bases.
Although the vast majority of knowledge bases KBs are heavily biased towards English, Wikipedias do cover very different topics in different languages. Exploiting this, we introduce a new multilingual dataset (X-WikiRE), framing relation extraction as a multilingual machine reading problem. We show that by leveraging this resource it is possible to robustly transfer models cross-lingually and that multilingual support significantly improves (zero-shot) relation extraction, enabling the population of low-resourced KBs from their well-populated counterparts.