Zero-shot Cross-Lingual Transfer for Synthetic Data Generation in Grammatical Error Detection

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Latouche, Gaetan Lopez, Carbonneau, Marc-André, Swanson, Ben
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917723870593024
author Latouche, Gaetan Lopez
Carbonneau, Marc-André
Swanson, Ben
author_facet Latouche, Gaetan Lopez
Carbonneau, Marc-André
Swanson, Ben
contents Grammatical Error Detection (GED) methods rely heavily on human annotated error corpora. However, these annotations are unavailable in many low-resource languages. In this paper, we investigate GED in this context. Leveraging the zero-shot cross-lingual transfer capabilities of multilingual pre-trained language models, we train a model using data from a diverse set of languages to generate synthetic errors in other languages. These synthetic error corpora are then used to train a GED model. Specifically we propose a two-stage fine-tuning pipeline where the GED model is first fine-tuned on multilingual synthetic data from target languages followed by fine-tuning on human-annotated GED corpora from source languages. This approach outperforms current state-of-the-art annotation-free GED methods. We also analyse the errors produced by our method and other strong baselines, finding that our approach produces errors that are more diverse and more similar to human errors.
format Preprint
id arxiv_https___arxiv_org_abs_2407_11854
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Zero-shot Cross-Lingual Transfer for Synthetic Data Generation in Grammatical Error Detection
Latouche, Gaetan Lopez
Carbonneau, Marc-André
Swanson, Ben
Computation and Language
Artificial Intelligence
Grammatical Error Detection (GED) methods rely heavily on human annotated error corpora. However, these annotations are unavailable in many low-resource languages. In this paper, we investigate GED in this context. Leveraging the zero-shot cross-lingual transfer capabilities of multilingual pre-trained language models, we train a model using data from a diverse set of languages to generate synthetic errors in other languages. These synthetic error corpora are then used to train a GED model. Specifically we propose a two-stage fine-tuning pipeline where the GED model is first fine-tuned on multilingual synthetic data from target languages followed by fine-tuning on human-annotated GED corpora from source languages. This approach outperforms current state-of-the-art annotation-free GED methods. We also analyse the errors produced by our method and other strong baselines, finding that our approach produces errors that are more diverse and more similar to human errors.
title Zero-shot Cross-Lingual Transfer for Synthetic Data Generation in Grammatical Error Detection
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2407.11854