{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,9]],"date-time":"2026-06-09T15:12:37Z","timestamp":1781017957338,"version":"3.54.1"},"reference-count":41,"publisher":"Association for Computing Machinery (ACM)","issue":"7","license":[{"start":{"date-parts":[[2023,7,20]],"date-time":"2023-07-20T00:00:00Z","timestamp":1689811200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Asian Low-Resour. Lang. Inf. Process."],"published-print":{"date-parts":[[2023,7,31]]},"abstract":"<jats:p>This article investigates approaches to effectively harness source-side linguistic features for low-resource multilingual neural machine translation (MNMT). Previous works focus on using various features of a word such as lemma, part-of-speech tag, dependency label, and so on, to improve translation quality in a low-resource scenario. However, these studies deal with bilingual translation and do not focus on using features in multilingual training setups. Our work focuses on this particular point and experiments with low-resource multilingual models incorporating source-side linguistic features. Although techniques for integrating features into an NMT model such as concatenation and feature relevance perform quite well in bilingual settings, they do not work well in multilingual settings. To remedy this, we propose the use of dummy features and language indicator features in MNMT models. Experiments are conducted on English to Asian language translation on a multilingual, multi-parallel corpus spanning English and eight Asian languages where for each language pair, the training data size does not exceed 20,000 parallel sentences. After establishing strong bilingual baselines using feature relevance mechanisms and multilingual baselines without any features, we show that our proposed dummy features and language indicator features, in combination with feature relevance mechanisms, yield significant improvements in BLEU points for all language pairs. We then analyze our models from the perspectives of model sizes, the impact of individual linguistic features, validation perplexity computed during training, visualization of the activations of the relevance mechanisms, and exhaustive tuning of hyperparameters. We also report preliminary results for multilingual multi-way models using linguistic features.<\/jats:p>","DOI":"10.1145\/3594631","type":"journal-article","created":{"date-parts":[[2023,5,18]],"date-time":"2023-05-18T12:17:16Z","timestamp":1684412236000},"page":"1-36","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":10,"title":["Low-resource Multilingual Neural Translation Using Linguistic Feature-based Relevance Mechanisms"],"prefix":"10.1145","volume":"22","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-5947-4775","authenticated-orcid":false,"given":"Abhisek","family":"Chakrabarty","sequence":"first","affiliation":[{"name":"NICT, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-0664-3421","authenticated-orcid":false,"given":"Raj","family":"Dabre","sequence":"additional","affiliation":[{"name":"NICT, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7523-208X","authenticated-orcid":false,"given":"Chenchen","family":"Ding","sequence":"additional","affiliation":[{"name":"NICT, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1111-9245","authenticated-orcid":false,"given":"Masao","family":"Utiyama","sequence":"additional","affiliation":[{"name":"NICT, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1028-4399","authenticated-orcid":false,"given":"Eiichiro","family":"Sumita","sequence":"additional","affiliation":[{"name":"NICT, Japan"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"320","published-online":{"date-parts":[[2023,7,20]]},"reference":[{"key":"e_1_3_3_2_2","unstructured":"Naveen Arivazhagan Ankur Bapna Orhan Firat Dmitry Lepikhin Melvin Johnson Maxim Krikun Mia Xu Chen Yuan Cao George Foster Colin Cherry Wolfgang Macherey Zhifeng Chen and Yonghui Wu. 2019. Massively multilingual neural machine translation in the wild: Findings and challenges. arxiv:1907.05019 [cs.CL]. Retrieved from https:\/\/arxiv.org\/abs\/1907.05019."},{"key":"e_1_3_3_3_2","volume-title":"Proceedings of the International Conference on Learning Representations","author":"Bahdanau Dzmitry","year":"2015","unstructured":"Dzmitry Bahdanau, Kyung Hyun Cho, and Yoshua Bengio. 2015. Neural machine translation by jointly learning to align and translate. In Proceedings of the International Conference on Learning Representations."},{"key":"e_1_3_3_4_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.acl-main.147"},{"key":"e_1_3_3_5_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W17-4703"},{"key":"e_1_3_3_6_2","doi-asserted-by":"crossref","first-page":"4263","DOI":"10.18653\/v1\/2020.coling-main.376","volume-title":"Proceedings of the 28th International Conference on Computational Linguistics","author":"Chakrabarty Abhisek","year":"2020","unstructured":"Abhisek Chakrabarty, Raj Dabre, Chenchen Ding, Masao Utiyama, and Eiichiro Sumita. 2020. Improving low-resource NMT through relevance based linguistic features incorporation. In Proceedings of the 28th International Conference on Computational Linguistics. International Committee on Computational Linguistics, 4263\u20134274."},{"key":"e_1_3_3_7_2","doi-asserted-by":"publisher","DOI":"10.1145\/2835494"},{"key":"e_1_3_3_8_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P17-1136"},{"key":"e_1_3_3_9_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P17-1177"},{"key":"e_1_3_3_10_2","doi-asserted-by":"publisher","DOI":"10.1145\/3406095"},{"key":"e_1_3_3_11_2","doi-asserted-by":"publisher","DOI":"10.1145\/3276773"},{"key":"e_1_3_3_12_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P16-1078"},{"key":"e_1_3_3_13_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/n16-1101"},{"key":"e_1_3_3_14_2","article-title":"Factored neural machine translation","author":"Garc\u00eda-Mart\u00ednez Mercedes","year":"2016","unstructured":"Mercedes Garc\u00eda-Mart\u00ednez, Lo\u00efc Barrault, and Fethi Bougares. 2016. Factored neural machine translation. arXiv:1609.04621. Retrieved from https:\/\/arxiv.org\/abs\/1609.04621.","journal-title":"arXiv:1609.04621"},{"key":"e_1_3_3_15_2","unstructured":"Mercedes Garc\u00eda-Mart\u00ednez Lo\u00efc Barrault and Fethi Bougares. 2016. Factored neural machine translation architectures. In Proceedings of the 13th International Conference on Spoken Language Translation International Workshop on Spoken Language Translation Seattle WA. Retrieved from https:\/\/aclanthology.org\/2016.iwslt-1.3."},{"key":"e_1_3_3_16_2","first-page":"7","volume-title":"Proceedings of the Australasian Language Technology Association Workshop 2016","author":"Hoang Cong Duy Vu","year":"2016","unstructured":"Cong Duy Vu Hoang, Reza Haffari, and Trevor Cohn. 2016. Improving neural translation models with linguistic factors. In Proceedings of the Australasian Language Technology Association Workshop 2016, 7\u201314."},{"key":"e_1_3_3_17_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W19-5319"},{"key":"e_1_3_3_18_2","first-page":"944","volume-title":"Proceedings of the Conference on Empirical Methods in Natural Language Processing","author":"Isozaki Hideki","year":"2010","unstructured":"Hideki Isozaki, Tsutomu Hirao, Kevin Duh, Katsuhito Sudoh, and Hajime Tsukada. 2010. Automatic evaluation of translation quality for distant language pairs. In Proceedings of the Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 944\u2013952."},{"key":"e_1_3_3_19_2","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00065"},{"key":"e_1_3_3_20_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P17-4012"},{"key":"e_1_3_3_21_2","doi-asserted-by":"publisher","DOI":"10.5555\/1734086"},{"key":"e_1_3_3_22_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P17-1064"},{"key":"e_1_3_3_23_2","doi-asserted-by":"publisher","DOI":"10.1109\/TASLP.2018.2864648"},{"key":"e_1_3_3_24_2","first-page":"55","volume-title":"Association for Computational Linguistics (ACL) System Demonstrations","author":"Manning Christopher D.","year":"2014","unstructured":"Christopher D. Manning, Mihai Surdeanu, John Bauer, Jenny Finkel, Steven J. Bethard, and David McClosky. 2014. The Stanford CoreNLP natural language processing toolkit. In Association for Computational Linguistics (ACL) System Demonstrations. 55\u201360."},{"key":"e_1_3_3_25_2","volume-title":"Advances in Neural Information Processing Systems","author":"Mikolov Tomas","year":"2013","unstructured":"Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S. Corrado, and Jeff Dean. 2013. Distributed representations of words and phrases and their compositionality. In Advances in Neural Information Processing Systems, C. J. C. Burges, L. Bottou, M. Welling, Z. Ghahramani, and K. Q. Weinberger (Eds.), Vol. 26. Curran Associates, Inc."},{"key":"e_1_3_3_26_2","doi-asserted-by":"publisher","DOI":"10.20944\/preprints202002.0273.v1"},{"key":"e_1_3_3_27_2","doi-asserted-by":"publisher","DOI":"10.3115\/1073083.1073135"},{"key":"e_1_3_3_28_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W15-3049"},{"key":"e_1_3_3_29_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W17-4770"},{"key":"e_1_3_3_30_2","first-page":"1","volume-title":"Proceedings of the Oriental COCOSDA International Conference on Speech Database and Assessments (O-COCOSDA\u201916)","author":"Riza Hammam","year":"2016","unstructured":"Hammam Riza, Michael Purwoadi, Gunarso, Teduh Uliniansyah, Aw Ai Ti, Sharifah Mahani Aljunied, Luong Chi Mai, Vu Tat Thang, Nguyen Phuong Thai, Rapid Sun, Vichet Chea, Khin Mar Soe, Khin Thandar Nwet, Masao Utiyama, and Chenchen Ding. 2016. Introduction of the Asian language treebank. In Proceedings of the Oriental COCOSDA International Conference on Speech Database and Assessments (O-COCOSDA\u201916). 1\u20136."},{"key":"e_1_3_3_31_2","doi-asserted-by":"publisher","DOI":"10.1007\/s10590-020-09258-6"},{"key":"e_1_3_3_32_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W16-2209"},{"key":"e_1_3_3_33_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P16-1162"},{"key":"e_1_3_3_34_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1021"},{"key":"e_1_3_3_35_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1159"},{"key":"e_1_3_3_36_2","first-page":"5926","volume-title":"International Conference on Machine Learning","author":"Song Kaitao","year":"2019","unstructured":"Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, and Tie-Yan Liu. 2019. MASS: Masked sequence to sequence pre-training for language generation. In International Conference on Machine Learning. 5926\u20135936."},{"key":"e_1_3_3_37_2","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00252"},{"key":"e_1_3_3_38_2","first-page":"5998","volume-title":"Advances in Neural Information Processing Systems 30","author":"Vaswani Ashish","year":"2017","unstructured":"Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, \u0141ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems 30, I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett (Eds.). Curran Associates, Inc., 5998\u20136008."},{"key":"e_1_3_3_39_2","doi-asserted-by":"publisher","DOI":"10.1109\/TASLP.2018.2855968"},{"key":"e_1_3_3_40_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.acl-main.148"},{"key":"e_1_3_3_41_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/N19-1118"},{"key":"e_1_3_3_42_2","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1163"}],"container-title":["ACM Transactions on Asian and Low-Resource Language Information Processing"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3594631","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3594631","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T18:09:08Z","timestamp":1750183748000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3594631"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,7,20]]},"references-count":41,"journal-issue":{"issue":"7","published-print":{"date-parts":[[2023,7,31]]}},"alternative-id":["10.1145\/3594631"],"URL":"https:\/\/doi.org\/10.1145\/3594631","relation":{},"ISSN":["2375-4699","2375-4702"],"issn-type":[{"value":"2375-4699","type":"print"},{"value":"2375-4702","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,7,20]]},"assertion":[{"value":"2021-03-19","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-04-11","order":1,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-07-20","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}