TreeSwap: Data Augmentation for Machine Translation via Dependency Subtree Swapping

Nagy, A and Lakatos, Dorina Petra and Barta, Botond and Ács, Judit (2023) TreeSwap: Data Augmentation for Machine Translation via Dependency Subtree Swapping. In: Large language models for natural language processing proceedings, RANLP 2023. ИНКОМА, Shumen, pp. 759-768. ISBN 9789544520922 10.26615/978-954-452-092-2_082

[img] Text
Nagy_759_34448299_ny.pdf

Download (386kB)

Abstract

Data augmentation methods for neural machine translation are particularly useful when limited amount of training data is available, which is often the case when dealing with low-resource languages. We introduce a novel augmentation method, which generates new sentences by swapping objects and subjects across bisentences. This is performed simultaneously based on the dependency parse trees of the source and target sentences. We name this method TreeSwap. Our results show that TreeSwap achieves consistent improvements over baseline models in 4 language pairs in both directions on resource-constrained datasets. We also explore domain-specific corpora, but find that our method does not make significant improvements on law, medical and IT data. We report the scores of similar augmentation methods and find that TreeSwap performs comparably. We also analyze the generated sentences qualitatively and find that the augmentation produces a correct translation in most cases. Our code is available on Github1. © 2023 Incoma Ltd. All rights reserved.

Item Type: Book Section
Uncontrolled Keywords: Computational linguistics; Training data; Dependency parser; Augmentation methods; Computer aided language translation; Data augmentation; neural machine translation; Machine translations; Low Resource Languages; Baseline models; Language pairs; Parse trees; Subtree swapping;
Subjects: Q Science > QA Mathematics and Computer Science > QA75 Electronic computers. Computer science / számítástechnika, számítógéptudomány
Divisions: Informatics Laboratory
Internet Applications Department
SWORD Depositor: MTMT Injector
Depositing User: MTMT Injector
Date Deposited: 17 Jan 2024 08:40
Last Modified: 17 Jan 2024 08:40
URI: https://eprints.sztaki.hu/id/eprint/10663

Update Item Update Item