Exploring Linguistic Properties of Monolingual BERTs with Typological Classification among Languages
This provides insights into transformer representations for NLP researchers, though it is incremental in analyzing existing models rather than proposing new methods.
The paper investigates how monolingual BERT models encode structural linguistic information by comparing weight matrices across typologically similar languages using Centered Kernel Alignment, finding that syntactic similarity correlates with weight similarity in middle layers and is enhanced by domain adaptation on semantically equivalent texts.
The impressive achievements of transformers force NLP researchers to delve into how these models represent the underlying structure of natural language. In this paper, we propose a novel standpoint to investigate the above issue: using typological similarities among languages to observe how their respective monolingual models encode structural information. We aim to layer-wise compare transformers for typologically similar languages to observe whether these similarities emerge for particular layers. For this investigation, we propose to use Centered Kernel Alignment to measure similarity among weight matrices. We found that syntactic typological similarity is consistent with the similarity between the weights in the middle layers, which are the pretrained BERT layers to which syntax encoding is generally attributed. Moreover, we observe that a domain adaptation on semantically equivalent texts enhances this similarity among weight matrices.