WikiSplit++: Easy Data Refinement for Split and Rephrase
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929313349107712 |
|---|---|
| author | Tsukagoshi, Hayato Hirao, Tsutomu Morishita, Makoto Chousa, Katsuki Sasano, Ryohei Takeda, Koichi |
| author_facet | Tsukagoshi, Hayato Hirao, Tsutomu Morishita, Makoto Chousa, Katsuki Sasano, Ryohei Takeda, Koichi |
| contents | The task of Split and Rephrase, which splits a complex sentence into multiple simple sentences with the same meaning, improves readability and enhances the performance of downstream tasks in natural language processing (NLP). However, while Split and Rephrase can be improved using a text-to-text generation approach that applies encoder-decoder models fine-tuned with a large-scale dataset, it still suffers from hallucinations and under-splitting. To address these issues, this paper presents a simple and strong data refinement approach. Here, we create WikiSplit++ by removing instances in WikiSplit where complex sentences do not entail at least one of the simpler sentences and reversing the order of reference simple sentences. Experimental results show that training with WikiSplit++ leads to better performance than training with WikiSplit, even with fewer training instances. In particular, our approach yields significant gains in the number of splits and the entailment ratio, a proxy for measuring hallucinations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_09002 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | WikiSplit++: Easy Data Refinement for Split and Rephrase Tsukagoshi, Hayato Hirao, Tsutomu Morishita, Makoto Chousa, Katsuki Sasano, Ryohei Takeda, Koichi Computation and Language The task of Split and Rephrase, which splits a complex sentence into multiple simple sentences with the same meaning, improves readability and enhances the performance of downstream tasks in natural language processing (NLP). However, while Split and Rephrase can be improved using a text-to-text generation approach that applies encoder-decoder models fine-tuned with a large-scale dataset, it still suffers from hallucinations and under-splitting. To address these issues, this paper presents a simple and strong data refinement approach. Here, we create WikiSplit++ by removing instances in WikiSplit where complex sentences do not entail at least one of the simpler sentences and reversing the order of reference simple sentences. Experimental results show that training with WikiSplit++ leads to better performance than training with WikiSplit, even with fewer training instances. In particular, our approach yields significant gains in the number of splits and the entailment ratio, a proxy for measuring hallucinations. |
| title | WikiSplit++: Easy Data Refinement for Split and Rephrase |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2404.09002 |