RuSentEval: Linguistic Source, Encoder Force!
This work addresses the gap in research on multilingual transformers for linguists and NLP researchers, though it is incremental as it extends existing probing methods to a new language.
The authors tackled the lack of understanding of linguistic knowledge in pre-trained transformers for non-English languages by introducing RuSentEval, a set of 14 probing tasks for Russian, and found that some linguistic properties are learned similarly across Russian and English, contradicting common assumptions.
The success of pre-trained transformer language models has brought a great deal of interest on how these models work, and what they learn about language. However, prior research in the field is mainly devoted to English, and little is known regarding other languages. To this end, we introduce RuSentEval, an enhanced set of 14 probing tasks for Russian, including ones that have not been explored yet. We apply a combination of complementary probing methods to explore the distribution of various linguistic properties in five multilingual transformers for two typologically contrasting languages -- Russian and English. Our results provide intriguing findings that contradict the common understanding of how linguistic knowledge is represented, and demonstrate that some properties are learned in a similar manner despite the language differences.