Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation
We present an easy and efficient method to extend existing sentence embedding models to new languages. This allows to create multilingual versions from previously monolingual models. The training is based on the idea that a translated sentence should be mapped to the same location in the vector spac...
Gespeichert in:
Hauptverfasser: | , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | We present an easy and efficient method to extend existing sentence embedding
models to new languages. This allows to create multilingual versions from
previously monolingual models. The training is based on the idea that a
translated sentence should be mapped to the same location in the vector space
as the original sentence. We use the original (monolingual) model to generate
sentence embeddings for the source language and then train a new system on
translated sentences to mimic the original model. Compared to other methods for
training multilingual sentence embeddings, this approach has several
advantages: It is easy to extend existing models with relatively few samples to
new languages, it is easier to ensure desired properties for the vector space,
and the hardware requirements for training is lower. We demonstrate the
effectiveness of our approach for 50+ languages from various language families.
Code to extend sentence embeddings models to more than 400 languages is
publicly available. |
---|---|
DOI: | 10.48550/arxiv.2004.09813 |