Assessing the Role of Lexical Semantics in Cross-lingual Transfer through Controlled Manipulations
This work addresses the challenge of cross-lingual transfer for NLP researchers, providing insights into key factors, though it is incremental as it builds on existing methods with controlled manipulations.
The paper tackled the problem of understanding the conditions for effective cross-lingual transfer by assessing the role of lexical semantics compared to other language properties, showing that lexical matching (measured by translation entropy) greatly affects alignment quality, while script or word order have limited impact.
While cross-linguistic model transfer is effective in many settings, there is still limited understanding of the conditions under which it works. In this paper, we focus on assessing the role of lexical semantics in cross-lingual transfer, as we compare its impact to that of other language properties. Examining each language property individually, we systematically analyze how differences between English and a target language influence the capacity to align the language with an English pretrained representation space. We do so by artificially manipulating the English sentences in ways that mimic specific characteristics of the target language, and reporting the effect of each manipulation on the quality of alignment with the representation space. We show that while properties such as the script or word order only have a limited impact on alignment quality, the degree of lexical matching between the two languages, which we define using a measure of translation entropy, greatly affects it.