Many-to-Many Multilingual Translation Model for Languages of Indonesia

被引:1
|
作者
Wongso, Wilson [1 ]
Joyoadikusumo, Ananto [1 ]
Buana, Brandon Scott [1 ]
Suhartono, Derwin [1 ]
机构
[1] Bina Nusantara Univ, Sch Comp Sci, Comp Sci Dept, Jakarta 11480, Indonesia
关键词
Languages of Indonesia; low-resource languages; mT5; natural language processing; neural machine translation; LOW-RESOURCE;
D O I
10.1109/ACCESS.2023.3308818
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Indonesia is home to over 700 languages and most people speak their respective regional languages aside from the lingua franca. In this paper, we focus on the task of multilingual machine translation for 45 regional Indonesian languages and introduced Indo-T5 which leveraged the mT5 sequence-to-sequence language model as a baseline. Performances of bilingual and multilingual fine-tuning methods were also compared, in which we found that our models have outperformed current state-of-the-art translation models. We also investigate the use of religious texts from the Bible as an intermediate mid-resource translation domain for low-resource translation domain specialization. Our findings suggest that this two-step fine-tuning approach is highly effective in improving the quality of translations for low-resource text domains. Our results show an increase in SacreBLEU scores when evaluated on the low-resource NusaX dataset. We release our translation models for other researchers to leverage.
引用
收藏
页码:91385 / 91397
页数:13
相关论文
共 50 条
  • [31] An optimization model and a solution algorithm for the many-to-many car pooling problem
    Yan, Shangyao
    Chen, Chun-Ying
    ANNALS OF OPERATIONS RESEARCH, 2011, 191 (01) : 37 - 71
  • [32] APPROXIMATE ANALYTIC MODEL OF MANY-TO-MANY DEMAND RESPONSIVE TRANSPORTATION SYSTEMS
    DAGANZO, CF
    TRANSPORTATION RESEARCH, 1978, 12 (05): : 325 - 333
  • [33] Improving Many-to-Many Neural Machine Translation via Selective and Aligned Online Data Augmentation
    Zhang, Weitai
    Dai, Lirong
    Liu, Junhua
    Wang, Shijin
    APPLIED SCIENCES-BASEL, 2023, 13 (06):
  • [34] Remote intelligence enables many-to-many solutions
    Condon, Michael
    Control Engineering, 2022, 69 (10)
  • [35] Many-to-Many Superpixel Matching for Robust Tracking
    Wang, Junqiu
    Yagi, Yasushi
    IEEE TRANSACTIONS ON CYBERNETICS, 2014, 44 (07) : 1237 - 1248
  • [36] On the scalability of many-to-many reliable multicast sessions
    Yoon, WY
    Lee, D
    Youn, HY
    CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2004, 16 (13): : 1353 - 1363
  • [37] SMART: A many-to-many multicast protocol for ATM
    Gauthier, E
    LeBoudec, JY
    Oechslin, P
    IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, 1997, 15 (03) : 458 - 472
  • [38] Popular critical matchings in the many-to-many setting
    Nasre, Meghana
    Nimbhorkar, Prajakta
    Ranjan, Keshav
    Sarkar, Ankita
    THEORETICAL COMPUTER SCIENCE, 2024, 982
  • [39] Pareto optimality in many-to-many matching problems
    Cechlarova, Katarina
    Eirinakis, Pavlos
    Fleiner, Tamas
    Magos, Dimitrios
    Mourtos, Ioannis
    Potpinkova, Eva
    DISCRETE OPTIMIZATION, 2014, 14 : 160 - 169
  • [40] Describing Behavior of Processes with Many-to-Many Interactions
    Fahland, Dirk
    APPLICATION AND THEORY OF PETRI NETS AND CONCURRENCY, PETRI NETS 2019, 2019, 11522 : 3 - 24