This repository contains information aimed at reproducing the experiments on multidomain machine translation described in two published papers:
- Minh Quang Pham, Josep Crego, François Yvon. Revisiting Multi-Domain Machine Translation, to appear in Transactions of the ACL, 20. See TACL2020 directory.
- Minh Quang Pham, Josep Crego, François Yvon, Jean Senellart. A Study of Residual Adapters for Multi-Domain Neural Machine Translation, to appear in the proceedings of the conference on Machine Translation (WMT) 2020. See TACL2020 directory.
References:
@article{pham-etal-2021-revisiting, title = "Revisiting Multi-Domain Machine Translation", author = "Pham, MinhQuang and Crego, Josep Maria and Yvon, Fran{\c{c}}ois", journal = "Transactions of the Association for Computational Linguistics", volume = "9", year = "2021", url = "https://aclanthology.org/2021.tacl-1.2", doi = "10.1162/tacl_a_00351", pages = "17--35", abstract = "When building machine translation systems, one often needs to make the best out of heterogeneous sets of parallel data in training, and to robustly handle inputs from unexpected domains in testing. This multi-domain scenario has attracted a lot of recent work that fall under the general umbrella of transfer learning. In this study, we revisit multi-domain machine translation, with the aim to formulate the motivations for developing such systems and the associated expectations with respect to performance. Our experiments with a large sample of multi-domain systems show that most of these expectations are hardly met and suggest that further work is needed to better analyze the current behaviour of multi-domain systems and to make them fully hold their promises.", }