Finding the Right Recipe for Low Resource Domain Adaptation in Neural Machine Translation
This work provides practical guidance for machine translation practitioners in specialized domains, but it is incremental as it focuses on empirical comparisons of existing methods.
The paper tackled the problem of low-resource domain adaptation in neural machine translation by empirically comparing data-centric methods under different data availability scenarios, finding that an ensemble approach improved in-domain performance while mitigating reductions in original domain quality across three domains and four language pairs.
General translation models often still struggle to generate accurate translations in specialized domains. To guide machine translation practitioners and characterize the effectiveness of domain adaptation methods under different data availability scenarios, we conduct an in-depth empirical exploration of monolingual and parallel data approaches to domain adaptation of pre-trained, third-party, NMT models in settings where architecture change is impractical. We compare data centric adaptation methods in isolation and combination. We study method effectiveness in very low resource (8k parallel examples) and moderately low resource (46k parallel examples) conditions and propose an ensemble approach to alleviate reductions in original domain translation quality. Our work includes three domains: consumer electronic, clinical, and biomedical and spans four language pairs - Zh-En, Ja-En, Es-En, and Ru-En. We also make concrete recommendations for achieving high in-domain performance and release our consumer electronic and medical domain datasets for all languages and make our code publicly available.