Assessing the Syntactic Capabilities of Transformer-based Multilingual Language Models
This work addresses the problem of understanding how multilingual language models handle syntactic structures across languages, which is important for researchers in NLP and linguistics, though it is incremental as it builds on existing evaluation frameworks.
The study assessed the syntactic generalization capabilities of monolingual and multilingual Transformer-based language models (BERT and RoBERTa) on English and Spanish, finding that multilingual models showed varied performance across languages, with specific comparisons made between monolingual and multilingual versions on English and between languages using introduced Spanish tests.
Multilingual Transformer-based language models, usually pretrained on more than 100 languages, have been shown to achieve outstanding results in a wide range of cross-lingual transfer tasks. However, it remains unknown whether the optimization for different languages conditions the capacity of the models to generalize over syntactic structures, and how languages with syntactic phenomena of different complexity are affected. In this work, we explore the syntactic generalization capabilities of the monolingual and multilingual versions of BERT and RoBERTa. More specifically, we evaluate the syntactic generalization potential of the models on English and Spanish tests, comparing the syntactic abilities of monolingual and multilingual models on the same language (English), and of multilingual models on two different languages (English and Spanish). For English, we use the available SyntaxGym test suite; for Spanish, we introduce SyntaxGymES, a novel ensemble of targeted syntactic tests in Spanish, designed to evaluate the syntactic generalization capabilities of language models through the SyntaxGym online platform.