Characterizing Departures from Linearity in Word Translation
arXiv:1806.04508
Abstract
We investigate the behavior of maps learned by machine translation methods. The maps translate words by projecting between word embedding spaces of different languages. We locally approximate these maps using linear maps, and find that they vary across the word embedding space. This demonstrates that the underlying maps are non-linear. Importantly, we show that the locally linear maps vary by an amount that is tightly correlated with the distance between the neighborhoods on which they are trained. Our results can be used to test non-linear methods, and to drive the design of more accurate maps for word translation.
arXiv commentACL 2018
References in corpus (8)
- From Frequency to Meaning: Vector Space Models of Semantics
- Exploiting Similarities among Languages for Machine Translation
- An Autoencoder Approach to Learning Bilingual Word Representations
- BilBOWA: Fast Bilingual Distributed Representations without Word Alignments
- Massively Multilingual Word Embeddings
- Cross-lingual Models of Word Embeddings: An Empirical Comparison
- Learning Bilingual Word Representations by Marginalizing Alignments
- Learning Translations via Matrix Completion