Towards Efficient Dialogue Processing in the Emergency Response Domain

Anikina, Tatiana

doi:10.18653/v1/2023.acl-srw.31

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Adapters represent a small amount of additional parameters that can be added as trainable task-specific weights at each layer of the transformer architecture (Vaswani et al, 2017). They have been successful on a variety tasks including speech recognition (Hou et al, 2021), cross-lingual transfer (Parovic et al, 2022) and classification tasks (Lee et al, 2022;Anikina, 2023;Metheniti et al, 2023) but there is very little research on using adapters for coreference resolution and the only work that we are aware of uses parallel data for training (Tang and Hardmeier, 2023).…”

Section: Related Workmentioning

confidence: 99%

Multilingual coreference resolution: Adapt and Generate

Skachkova,

Anikina,

Mokhova

2023

Proceedings of the CRAC 2023 Shared Task on Multilingual Coreference Resolution

View full text Add to dashboard Cite

The paper presents two multilingual coreference resolution systems submitted for the CRAC Shared Task 2023. The DFKI-Adapt system achieves 61.86 F1 score on the shared task test data, outperforming the official baseline by 4.9 F1 points. This system uses a combination of different features and training settings, including character embeddings, adapter modules, joint pre-training and loss-based retraining. We provide evaluation for each of the settings on 12 different datasets and compare the results. The other submission DFKI-MPrompt uses a novel approach that involves prompting for mention generation. Although the scores achieved by this model are lower compared to the baseline, the method shows a new way of approaching the coreference task and provides good results with just five epochs of training.

show abstract