DiscoFlan: Instruction Fine-tuning and Refined Text Generation for Discourse Relation Label Classification

Kaveri Anuranjana


Abstract
This paper introduces DiscoFlan, a multilingual discourse relation classifier submitted for DISRPT 2023. Our submission represents the first attempt at building a multilingual discourse relation classifier for the DISRPT 2023 shared task. By our model addresses the issue to mismatches caused by hallucination in a seq2seq model by utilizing the label distribution information for label generation. In contrast to the previous state-of-the-art model, our approach eliminates the need for hand-crafted features in computing the discourse relation classes. Furthermore, we propose a novel label generation mechanism that anchors the labels to a fixed set by selectively enhancing training on the decoder model. Our experimental results demonstrate that our model surpasses the current state-of-the-art performance in 11 out of the 26 datasets considered, however the submitted model compatible with provided evaluation scripts is better in 7 out of 26 considered datasets, while demonstrating competitive results in the rest. Overall, DiscoFlan showcases promising advancements in multilingual discourse relation classification for the DISRPT 2023 shared task.
Anthology ID:
2023.disrpt-1.2
Volume:
Proceedings of the 3rd Shared Task on Discourse Relation Parsing and Treebanking (DISRPT 2023)
Month:
July
Year:
2023
Address:
Toronto, Canada
Editors:
Chloé Braud, Yang Janet Liu, Eleni Metheniti, Philippe Muller, Laura Rivière, Attapol Rutherford, Amir Zeldes
Venue:
DISRPT
SIG:
Publisher:
The Association for Computational Linguistics
Note:
Pages:
22–28
Language:
URL:
https://aclanthology.org/2023.disrpt-1.2
DOI:
10.18653/v1/2023.disrpt-1.2
Bibkey:
Cite (ACL):
Kaveri Anuranjana. 2023. DiscoFlan: Instruction Fine-tuning and Refined Text Generation for Discourse Relation Label Classification. In Proceedings of the 3rd Shared Task on Discourse Relation Parsing and Treebanking (DISRPT 2023), pages 22–28, Toronto, Canada. The Association for Computational Linguistics.
Cite (Informal):
DiscoFlan: Instruction Fine-tuning and Refined Text Generation for Discourse Relation Label Classification (Anuranjana, DISRPT 2023)
Copy Citation:
PDF:
https://aclanthology.org/2023.disrpt-1.2.pdf