Neural Machine Translation Models with Back-Translation for the Extremely Low-Resource Indigenous Language Bribri

Proceedings of the 28th International Conference on Computational Linguistics, pages 3965–3976 Barcelona, Spain (Online), December 8-13, 2020 3965 Neural Machine Translation Models with Back-Translation for the Extremely Low-Resource Indigenous Language Bribri Isaac Feldman Dartmouth College isaac.c.feldman.23@dartmouth.edu Rolando Coto-Solano Dartmouth College rolando.a.coto.solano@dartmouth.edu Abstract This paper presents a neural machine translation model and dataset for the Chibchan language Bribri, with an average performance of BLEU 16.9±1.7. This was trained on an extremely small dataset (5923 Bribri-Spanish pairs), providing evidence for the applicability of NMT in extremely low-resource environments. We discuss the challenges entailed in managing training input from languages without standard orthographies, we provide evidence of successful learning of Bribri grammar, and also examine the translations of structures that are infrequent in major Indo-European languages, such as positional verbs, ergative markers, numerical classifiers and complex demonstrative systems. In addition to this, we perform an experiment of augmenting the dataset through iterative back-translation (Sennrich et al., 2016a; Hoang et al., 2018), by using Spanish sentences to create synthetic Bribri sentences. This improves the score by an average of 1.0 BLEU, but only when the new Spanish sentences belong to the same domain as the other Spanish examples. This contributes to the small but growing body of research on Chibchan NLP. 1 Introduction State-of-the-art systems in neural machine translation require large amounts of rich and varied parallel data between the source and target language (Lui et al., 2019). While this is less of a problem for high- resource language pairs like English-German or English-Spanish, there are much fewer resources when one of the languages is an Indigenous language. While there are some researchers creating systems to optimize the training of NMT frameworks with lower resources, most work primarily from a “warm- start” where models are pre-trained on a small, but still significant amount of parallel data (He et al., 2016). For example even in papers geared towards developing techniques for low resource languages, training datasets can exceed millions of sentences (Edunov et al., 2018). Our motivation in adapting low- resource techniques to extremely low-resource situations is that they can help improve understanding of their underlying mechanisms and lower the barrier of entry for machine translation to be used for language revitalization purposes. However, in the effective body of research, many papers use simulated low-resource scenarios where high resource parallel datasets are artificially truncated to test the efficacy of their techniques. While this does provide a level of standardization across different techniques, it is still artificial. This paper seeks to explore NMT in an authentic scenario, using an actual extremely low-resource language. We leverage a variety of sources to produce a novel translation dataset in the indigenous Bribri lan- guage for demonstrating methods of neural machine translation (NMT) in extremely low-resource sit- uations. To demonstrate usage of the dataset, we apply the technique of iterative back-translation with validation (Hoang et al., 2018). Finally, we present a translation analysis to show the unique challenges that this Bribri dataset presents. This work is licensed under a Creative …

Content truncated to 500 words