Machine Translation Pre-training for Data-to-Text Generation -- A Case Study in Czech
This work addresses the lack of research in data-to-text generation for non-English languages, offering a practical solution for Czech, though it is incremental as it adapts existing pre-training methods to a new domain.
The paper tackled the problem of data-to-text generation for non-English languages, specifically Czech, by using machine translation pre-training, resulting in significantly improved performance in automatic metrics and human evaluation, with benefits like better low-data performance and robustness to unseen slot values.
While there is a large body of research studying deep learning methods for text generation from structured data, almost all of it focuses purely on English. In this paper, we study the effectiveness of machine translation based pre-training for data-to-text generation in non-English languages. Since the structured data is generally expressed in English, text generation into other languages involves elements of translation, transliteration and copying - elements already encoded in neural machine translation systems. Moreover, since data-to-text corpora are typically small, this task can benefit greatly from pre-training. Based on our experiments on Czech, a morphologically complex language, we find that pre-training lets us train end-to-end models with significantly improved performance, as judged by automatic metrics and human evaluation. We also show that this approach enjoys several desirable properties, including improved performance in low data scenarios and robustness to unseen slot values.