RST Discourse Parsing as Text-to-Text Generation

Hu, Xinyu; Wan, Xiaojun

doi:10.1109/taslp.2023.3306710

Cited by 1 publication

(3 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, most of them split the parsing process into two steps: EDU segmentation and RST tree prediction, for which the gold EDU labels are often required. Considering that the datasets of DocMT are not equipped with such information, we follow Hu and Wan (2023) to train an end-to-end RST parser from scratch through a Seq2Seq reformulation method.…”

Section: Rst Parsingmentioning

confidence: 99%

“…The linearized sequence is designed to contain the complete original input text for better performance, according to the observation of Paolini et al (2021). More details can be found in Hu and Wan (2023), and an example is shown in Figure 3(d).…”

Section: Rst Parsingmentioning

confidence: 99%

“…We employ the end-to-end method of Hu and Wan (2023) to train our RST parsing model. The parsing task is reformulated into a Seq2Seq task through a linearization process and then trained based on a pretrained language model.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Exploring Discourse Structure in Document-level Machine Translation

Hu,

Wan

2023

Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

Neural machine translation has achieved great success in the past few years with the help of transformer architectures and large-scale bilingual corpora. However, when the source text gradually grows into an entire document, the performance of current methods for documentlevel machine translation (DocMT) is less satisfactory. Although the context is beneficial to the translation in general, it is difficult for traditional methods to utilize such long-range information. Previous studies on DocMT have concentrated on extra contents such as multiple surrounding sentences and input instances divided by a fixed length. We suppose that they ignore the structure inside the source text, which leads to under-utilization of the context. In this paper, we present a more sound paragraph-to-paragraph translation mode and explore whether discourse structure can improve DocMT. We introduce several methods from different perspectives, among which our RST-Att model with a multi-granularity attention mechanism based on the RST parsing tree works best. The experiments show that our method indeed utilizes discourse information and performs better than previous work.

show abstract

Section: Rst Parsingmentioning

confidence: 99%

Section: Rst Parsingmentioning

confidence: 99%