A Massively Multilingual Analysis of Cross-linguality in Shared Embedding Space

TL;DR

This study investigates factors influencing cross-lingual sentence alignment in shared embedding spaces across 101 languages, revealing key linguistic predictors and the impact of training data composition using large multilingual models.

Contribution

It provides a comprehensive analysis of linguistic and training factors affecting cross-lingual alignment in shared embedding spaces for numerous languages.

Findings

Word order agreement strongly predicts cross-lingual alignment.

Agreement in morphological complexity influences alignment quality.

In-family training data outperforms language-specific data in alignment metrics.

Abstract

In cross-lingual language models, representations for many different languages live in the same space. Here, we investigate the linguistic and non-linguistic factors affecting sentence-level alignment in cross-lingual pretrained language models for 101 languages and 5,050 language pairs. Using BERT-based LaBSE and BiLSTM-based LASER as our models, and the Bible as our corpus, we compute a task-based measure of cross-lingual alignment in the form of bitext retrieval performance, as well as four intrinsic measures of vector space alignment and isomorphism. We then examine a range of linguistic, quasi-linguistic, and training-related features as potential predictors of these alignment metrics. The results of our analyses show that word order agreement and agreement in morphological complexity are two of the strongest linguistic predictors of cross-linguality. We also note in-family training data as a stronger predictor than language-specific training data across the board. We verify some of our linguistic findings by looking at the effect of morphological segmentation on English-Inuktitut alignment, in addition to examining the effect of word order agreement on isomorphism for 66 zero-shot language pairs from a different corpus. We make the data and code for our experiments publicly available.