Analogy Training Multilingual Encoderss
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Dokumenter
- Analogy Training Multilingual Encoders∗
Forlagets udgivne version, 279 KB, PDF-dokument
Language encoders encode words and phrases in ways that capture their local semantic relatedness, but are known to be globally inconsistent. Global inconsistency can seemingly be corrected for, in part, by leveraging signals from knowledge bases, but previous results are partial and limited to monolingual English encoders. We extract a large-scale multilingual, multi-word analogy dataset from Wikidata for diagnosing and correcting for global inconsistencies, and then implement a four-way Siamese BERT architecture for grounding multilingual BERT (mBERT) in Wikidata through analogy training. We show that analogy training not only improves the global consistency of mBERT, as well as the isomorphism of language-specific subspaces, but also leads to consistent gains on downstream tasks such as bilingual dictionary induction and sentence retrieval.
Originalsprog | Engelsk |
---|---|
Titel | Proceedings of the AAAI-21 International Joint Conference on Artificial Intelligence |
Antal sider | 10 |
Forlag | AAAI Press |
Publikationsdato | 2021 |
Sider | 12884-12892. |
ISBN (Elektronisk) | 978-1-57735-866-4 |
Status | Udgivet - 2021 |
Begivenhed | 35th AAAI Conference on Artificial Intelligence - Virtual Varighed: 2 feb. 2021 → 9 feb. 2021 |
Konference
Konference | 35th AAAI Conference on Artificial Intelligence |
---|---|
By | Virtual |
Periode | 02/02/2021 → 09/02/2021 |
Navn | Proceedings of the International Joint Conference on Artificial Intelligence |
---|---|
Nummer | 14 |
Vol/bind | 35 |
ISSN | 1045-0823 |
Antal downloads er baseret på statistik fra Google Scholar og www.ku.dk
Ingen data tilgængelig
ID: 300671526