Beyond English-Only Reading Comprehension: Experiments in Zero-Shot Multilingual Transfer for Bulgarian

TL;DR

This paper explores zero-shot multilingual transfer for reading comprehension by fine-tuning multilingual BERT on English datasets and applying it to Bulgarian, introducing a new dataset and achieving promising results in non-English language understanding.

Contribution

It demonstrates the effectiveness of multilingual BERT in zero-shot transfer for Bulgarian reading comprehension and introduces a new Bulgarian dataset with knowledge retrieval strategies.

Findings

Multilingual BERT achieves 42.23% accuracy on Bulgarian reading comprehension.

Knowledge retrieval from Wikipedia improves model performance.

Zero-shot transfer outperforms baseline methods.

Abstract

Recently, reading comprehension models achieved near-human performance on large-scale datasets such as SQuAD, CoQA, MS Macro, RACE, etc. This is largely due to the release of pre-trained contextualized representations such as BERT and ELMo, which can be fine-tuned for the target task. Despite those advances and the creation of more challenging datasets, most of the work is still done for English. Here, we study the effectiveness of multilingual BERT fine-tuned on large-scale English datasets for reading comprehension (e.g., for RACE), and we apply it to Bulgarian multiple-choice reading comprehension. We propose a new dataset containing 2,221 questions from matriculation exams for twelfth grade in various subjects -history, biology, geography and philosophy-, and 412 additional questions from online quizzes in history. While the quiz authors gave no relevant context, we incorporate knowledge from Wikipedia, retrieving documents matching the combination of question + each answer option. Moreover, we experiment with different indexing and pre-training strategies. The evaluation results show accuracy of 42.23%, which is well above the baseline of 24.89%.