Learning a Multi-Domain Curriculum for Neural Machine Translation
This work addresses the challenge of balancing and enhancing translation quality across multiple domains, which is incremental as it builds on existing data selection research.
The paper tackled the problem of data selection for multiple domains in neural machine translation by introducing domain-relevance features and a training curriculum, resulting in simultaneous performance improvements across domains, including out-of-domain, with solid gains over non-curriculum methods.
Most data selection research in machine translation focuses on improving a single domain. We perform data selection for multiple domains at once. This is achieved by carefully introducing instance-level domain-relevance features and automatically constructing a training curriculum to gradually concentrate on multi-domain relevant and noise-reduced data batches. Both the choice of features and the use of curriculum are crucial for balancing and improving all domains, including out-of-domain. In large-scale experiments, the multi-domain curriculum simultaneously reaches or outperforms the individual performance and brings solid gains over no-curriculum training.