Testing pre-trained Transformer models for Lithuanian news clustering
This work addresses clustering for Lithuanian news, an incremental improvement for a low-resource language domain.
The study compared pre-trained multilingual Transformer models (BERT, XLM-R) with older methods for Lithuanian news clustering, finding that fine-tuned Transformers outperform word vectors but underperform doc2vec embeddings.
A recent introduction of Transformer deep learning architecture made breakthroughs in various natural language processing tasks. However, non-English languages could not leverage such new opportunities with the English text pre-trained models. This changed with research focusing on multilingual models, where less-spoken languages are the main beneficiaries. We compare pre-trained multilingual BERT, XLM-R, and older learned text representation methods as encodings for the task of Lithuanian news clustering. Our results indicate that publicly available pre-trained multilingual Transformer models can be fine-tuned to surpass word vectors but still score much lower than specially trained doc2vec embeddings.