Pith. sign in

REVIEW 4 major objections 6 minor 43 references

Low-resource Machine Translation: what for? who for? An observational study on a dedicated Tetun language translation service

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An analysis of 100,000 real translation requests shows Tetun users mostly translate short educational texts from high-resource languages into Tetun, clashing with the news-dominated corpora used to train low-resource MT systems.

desk verdict First large-scale observational evidence on who uses low-resource MT; the domain-mismatch claim is plausible and partly corroborated, but the classifier needs stronger validation. read the letter →

arxiv 2411.12262 v4 pith:EN36TCXY submitted 2024-11-19 cs.CL cs.AI

Raphael Merx , Adérito José Guterres Correia , Hanna Suominen , Ekaterina Vylomova This is my paper · ORCID
classification cs.CLcs.AI
keywords machinetranslationlow-resourcelanguagesTetunobservationalstudydomainmismatchuserbehaviorcorpusanalysisTimor-Leste
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that machine translation for the low-resource, institutionalized language Tetun is being built for the wrong purpose. By analyzing 100,000 real requests to the tetun.org service, it finds that users, many of them students on mobile devices, mostly translate short texts from English, Portuguese, or Indonesian into Tetun, with heavy concentration in science, health, and education. Available Tetun corpora, by contrast, are dominated by news about government and law. The paper concludes that MT systems for Tetun should be optimized for the high-resource-to-Tetun direction, for educational domains, and for short inputs, and that evaluation should reflect these real usage patterns.

What carries the argument

The argument is carried by an observational pipeline that turns server logs into domain statistics. Topic and provenance labels are assigned by a prompt-based large language model (the paper uses Llama 3.1 8B) applied to the high-resource side of each translation pair, after LDA topic discovery helps define the label set; the same classifier is applied to the MADLAD-400 Tetun corpus after machine-translating it to English. Word-count weighting keeps short inputs from being underrepresented. A secondary mechanism is the perplexity comparison: a monolingual Tetun language model trained mostly on MADLAD gives average perplexity 1,774 on real MT inputs versus 153 on an unseen held-out corpus, and lower perplexity on the same news/government topics that dominate training data, tying the corpus mismatch to concrete modeling cost.

What would settle it

Manually annotate a random sample of, say, 500 user requests and 500 MADLAD documents with the same label set, compute the domain gap using the manual labels instead of the classifier's, and check whether the gap between user text and corpus narrows; if it drops below the reported levels, the paper's mismatch claim fails. A cheaper check is to re-run the classification pipeline feeding the Tetun side through a Tetun-capable classifier, if one is available, and compare the resulting distributions.

Watch

Extended reading notes

Core claim

The paper's central claim is a domain mismatch between actual translation demand and available training data, quantified on 100,000 server logs from March-August 2024. When weighted by word count, 'Science & research' appears in 34.2% of MT inputs but only 3.2% of the MADLAD-400 Tetun corpus, and 'Education & research material' appears in 55.9% of inputs versus 7.5% of MADLAD, while news articles and government topics dominate the corpus at 61.2% and 83.1% respectively. Users also translate into Tetun about 70% of the time and strongly prefer short texts, with a median input of 8 words. The paper interprets these numbers as evidence that the high-resource-to-low-resource direction and education-centric domains are what a low-resource MT service actually serves, and that current corpora misrepresent that demand.

Load-bearing premise

The central claim depends on the assumption that the classifier's errors are identical on user requests and on the machine-translated comparison corpus, since the classifier only reaches weighted F1 of 0.76 (topic) and 0.73 (provenance) and its error rates on the two pipelines were never compared.

Editorial extensions

If this is right

  • MT evaluation for Tetun should be anchored in educational, science, and health domains rather than news and government, because those are what users actually translate.
  • The high-resource-to-Tetun direction should receive priority in model development and corpus collection, since it accounts for over 70% of requests.
  • Short inputs, including single words, are the typical case with a median of 8 words, so evaluation should include them rather than focusing on full sentences and paragraphs.
  • Perplexity gaps indicate that current corpora leave underrepresented domains poorly modeled, so corpus collection should intentionally cover education, science, and health material.
  • Mobile-first inference matters because over 90% of usage is on mobile devices, particularly Android.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same observational approach could be applied to other institutionalized low-resource languages with a public translation service, providing a usage-grounded alternative to surveys and focus groups.
  • If the domain mismatch generalizes, it suggests that 'low-resource' corpus-building efforts in other settings may be systematically misaligned with actual community demand, and that cheap observational log analysis should precede corpus collection.
  • The perplexity numbers hint that a language model fine-tuned on user-like educational inputs could substantially reduce error on the real workload, a testable extension the paper does not run.
  • The paper's finding that users translate from Tetun mostly for daily-life and personal communication suggests that outward-facing communication is a separate use case requiring different optimization from the dominant inward direction.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper presents an observational study of 100,000 translation requests logged by tetun.org, a dedicated Tetun translation service. Using server logs and analytics data, it characterizes users (mostly mobile users in Timor-Leste, plausibly students), translation directions (mostly high-resource into Tetun), input lengths (mostly short texts), and domains. Domain is decomposed into topic and provenance, and labels are produced by Llama 3.1 8B applied to the high-resource side of each translation pair, with a weighted F1 of 0.76 for topic and 0.73 for provenance on a 100-log, single-annotator test set. These labels are compared against the MADLAD-400 Tetun corpus, whose texts are first machine-translated to English with MADLAD-3B and then classified with the same model. The paper reports a large mismatch: Tetun.org inputs are dominated by science/research and education/research material, while MADLAD is dominated by government and news provenance. The paper concludes with recommendations to prioritize the high-resource-to-Tetun direction, educational/scientific domains, short texts, and mobile-friendly MT.

Significance. If the reported domain mismatch is real, this is a valuable contribution: it is among the first large-scale observational studies of actual end-user MT behavior for a low-resource language, and it provides concrete evidence that corpus-driven assumptions about domain coverage can diverge sharply from community needs. The paper also ships a useful artifact in the form of a machine-translated English version of the MADLAD Tetun corpus, and it is careful to state limitations about selection bias and generalizability. The recommendations for evaluation design and corpus collection are actionable. However, the strength of the empirical conclusion currently rests on a classifier validation that does not cover the MADLAD-side pipeline, and the uncertainty in the headline percentages is not quantified. These gaps are substantial enough that the central claim, while plausible and well-motivated, is not yet established to the standard the recommendations require.

major comments (4)
  1. [§3.1.1, §3.1.4, Tables 3 and 4] The headline domain-mismatch claim is load-bearing, but the MADLAD-side labels are never validated. The classifier's weighted F1 (0.76 topic, 0.73 provenance) is measured only on a 100-log test set of user requests, whereas for MADLAD the same classifier is applied to English translations produced by a different model (MADLAD-3B). The paper does not report per-class precision/recall, nor does it show that classification error is comparable across the two pipelines. If, for example, the classifier over-tags short instructional user texts as 'Science & research' or 'Education & research material' and over-tags translated news articles as 'Government & socio-economic issues' or 'News article & press release', the 10-fold gaps in Tables 3 and 4 could be substantially artifactual. I ask the authors to validate the classifier on a sample of MADLAD documents (after their English translation) and to report a confusion matrix or per-class precision/recall for both pipelines, or otherwise bound the differential error.
  2. [§3.1.1, §4.3] Because domain percentages are multi-label and word-weighted, the paper should report uncertainty. No confidence intervals or significance tests are given for the percentages in Tables 3 and 4, and several adjacent categories are close (e.g., Education at 22.8% vs. Business & work & employment at 21.2%). A bootstrap over documents, or an inter-annotator agreement measure on the 100-log test set, would clarify whether the reported ordering and the 'large discrepancy' with MADLAD are robust. This is especially important given that the classifier itself has only 0.76/0.73 weighted F1.
  3. [§3.1.1, §4.2] The high-resource side of the input-output pair is not symmetric across translation directions. For English/Indonesian/Portuguese-to-Tetun requests, the classifier sees the user's source text; for Tetun-to-English/Indonesian/Portuguese requests, it sees the MT output. The paper does not report classification accuracy separately by direction, although Table 2 shows that roughly 29% of requests have Tetun as the source. Translation errors in the high-resource output could systematically affect labels, especially for short inputs. I request a direction-wise breakdown of classifier performance or an explicit analysis of whether label distributions differ by which side of the pair is the high-resource text.
  4. [§4.4, §4.2] The perplexity comparison between MT inputs and the Labadain corpus is confounded by input length and document segmentation. MT inputs have a median of 8 words, and perplexity on very short fragments is typically inflated relative to longer sentences or documents. The paper's claim that MT inputs 'may contain vocabulary, structures, or topics that are underrepresented in scraped monolingual corpora' should be supported by a length-controlled comparison, for example by computing perplexity on matched-length bins or by reporting perplexity for MT inputs and Labadain sentences in the same length range.
minor comments (6)
  1. [Table 1] There are formatting issues in the table, such as '0 .19' and '0 .40' with an extra space; please fix the spacing.
  2. [§3.1.1] 'LLama 3.1 8B' should be 'Llama 3.1 8B' for consistency with the reference list and standard naming.
  3. [Figure 4] The y-axis tick labels ('0 k', '20 k', etc.) are difficult to read and the figure would benefit from explicit numeric tick values.
  4. [§3.1.4] The statement 'Upon acceptance, we will release our translation' should be rephrased to 'Upon publication' or made unconditional, since the artifact is already described as part of the paper's contribution.
  5. [Tables 3 and 4] The paper should clarify explicitly that percentages are word-weighted and multi-label, and that columns therefore need not sum to 100%; the current tables are easy to misread as exclusive categories.
  6. [§5] The recommendation to prioritize the 'high-resource to low-resource direction' would benefit from specifying which high-resource language(s) are most relevant, since English, Portuguese, and Indonesian differ in their user shares and in their distance to Tetun.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper is an observational measurement whose claims are not derived from their own inputs by construction.

full rationale

This paper reports an observational study of tetun.org server logs. Its central claims—user demographics, translation direction, text length, and domain distributions—are measured directly from the logs and from external corpora (MADLAD-400, Labadain), not derived from a fitted model whose parameters encode the conclusion. The domain classifier (Llama 3.1 8B) is selected by comparing several classifiers on an independently hand-annotated 100-log test set (Section 3.1.1), and the same classifier is then applied to both MT inputs and MADLAD documents. This creates a risk of measurement error or differential label bias across pipelines, but it is not circularity: the headline percentages in Tables 3 and 4 are not fitted to the paper's recommendation, and no parameter is defined in terms of the target conclusion. Perplexity comparisons (Section 4.4) use Goldfish, a model trained on MADLAD, and compare MT inputs to Labadain; this is an external benchmark, not a self-referential loop. The paper's self-citations (e.g., Merx et al., 2024; de Jesus and Nunes, 2024c) are contextual and corroborative, not load-bearing premises imported to force a conclusion. The acknowledged limitations—classifier accuracy, single-annotator test set, server-log selection bias—are validity concerns, not instances where an output equals an input by construction. No quoted step exhibits the specific reduction required to establish circularity.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim is an empirical measurement, so the ledger contains no fitted parameters for the headline result. The free parameters listed are preprocessing and topic-modeling thresholds that shape the label set. The axioms are the unvalidated premises on which the domain-mismatch and perplexity conclusions rest: cross-lingual classification fidelity, comparable classifier error on user text and translated corpora, representativeness of MADLAD, successful-request logs as a proxy for demand, and perplexity comparability despite text-length and spelling differences.

free parameters (6)
  • LDA number of topics = 15
    Selected by maximizing topic coherence in hyperparameter search (Section 3.1.1); shapes the topic taxonomy used to build the classification label set.
  • LDA passes = 10
    Chosen in the same coherence search (Section 3.1.1); part of the topic discovery pipeline.
  • LDA alpha = asymmetric
    Chosen in the same coherence search (Section 3.1.1).
  • Minimum document length for LDA = >5 words
    Documents of 5 words or fewer are excluded from topic discovery because they lack context (Section 3.1.1).
  • Minimum bigram frequency = 5 occurrences
    Bigrams appearing fewer than 5 times are dropped in the LDA preprocessing (Section 3.1.1).
  • Spelling-guide inclusion threshold = 20%
    Inputs where fewer than 20% of words are in the official spelling guide are excluded from genre analysis as almost certainly not Tetun (Section 3.1.3).
assumptions (5)
  • domain assumption Domain can be decomposed into topic, provenance, and genre following Saunders (2022).
    The entire analysis is structured around this decomposition, introduced in Section 3.1.
  • domain assumption The high-resource side of each translation pair faithfully represents the topic and provenance of the original request, and the Llama 3.1 8B classifier has comparable accuracy on user text and on machine-translated MADLAD text.
    Tetun is not supported by the classifier, so all domain labels come from the English, Indonesian, or Portuguese side; differential error between the two text types would corrupt the domain comparison in Tables 3 and 4.
  • domain assumption MADLAD-400 Tetun clean split is representative of available Tetun corpora for the domain comparison.
    The topic, provenance, and genre comparisons use only this corpus; a different corpus choice could change the reported mismatch.
  • domain assumption Server logs of successful translation requests are a valid proxy for user demand.
    The paper acknowledges in Limitations that only successful requests are logged and that users may avoid poor-performing domains, yet interprets request distributions as reflecting end-user need.
  • domain assumption Perplexity differences between MT inputs and Labadain reflect domain and vocabulary mismatch rather than text length, spelling, or noise differences.
    MT inputs are much shorter and noisier (median 8 words, median 71% official spelling) than corpus sentences, and no matching on length or spelling is applied in Section 4.4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Low-resource Machine Translation: what for? who for? An observational study on a dedicated Tetun language translation service." pith.science (2026). https://pith.science/paper/EN36TCXY

@misc{pith2026241112262,
  author       = {Pith},
  title        = {Pith review of: Low-resource Machine Translation: what for? who for? An observational study on a dedicated Tetun language translation service},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EN36TCXY}},
  note         = {Machine review of arXiv:2411.12262}
}
abstract

Low-resource machine translation (MT) presents a diversity of community needs and application challenges that remain poorly understood. To complement surveys and focus groups, which tend to rely on small samples of respondents, we propose an observational study on actual usage patterns of tetun$.$org, a specialized MT service for the Tetun language, which is the lingua franca in Timor-Leste. Our analysis of 100,000 translation requests reveals patterns that challenge assumptions based on existing corpora. We find that users, many of them students on mobile devices, typically translate text from a high-resource language into Tetun across diverse domains including science, healthcare, and daily life. This contrasts sharply with available Tetun corpora, which are dominated by news articles covering government and social issues. Our results suggest that MT systems for institutionalized minority languages like Tetun should prioritize accuracy on domains relevant to educational contexts, in the high-resource to low-resource direction. More broadly, this study demonstrates how observational analysis can inform low-resource language technology development, by grounding research in practical community needs.

Figures

Figures reproduced from arXiv: 2411.12262 by the authors.

Figure 1
Figure 1. Overview of our approach and research ques [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Monthly active users by device and service. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Number of requests by day / hour, in Asia/Dili (UTC+9). We observe a spike in the evenings before [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Histogram showing the frequency distribution [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Histograms of MT Tetun inputs and MAD￾LAD Tetun texts by fraction of words in the official INL spelling guide. (55.9%). “Organizational & formal documents” comes as a distant second at 18%. This again stands in contrast with monolingual corpora, where over 61% of docum…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 26 canonical work pages

  1. [1]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Steven Bird and Dean Yibarbuk. 2024. https://aclanthology.org/2024.eacl-long.50 Centering the Speech Community . In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics ( Volume 1: Long Papers ) , pages 826--839, St. Julian's, Malta. Association for Computational Linguistics

    work page 2024

  4. [4]

    Verena Blaschke, Christoph Purschke, Hinrich Schuetze, and Barbara Plank. 2024. https://aclanthology.org/2024.acl-short.74 What Do Dialect Speakers Want ? A Survey of Attitudes Towards Language Technology for German Dialects . In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics ( Volume 2: Short Papers ) , pages 823-...

    work page 2024

  5. [5]

    Blei, Andrew Y

    David M. Blei, Andrew Y. Ng, and Michael I. Jordan. 2003. Latent dirichlet allocation. J. Mach. Learn. Res., 3:993--1022

    work page 2003

  6. [6]

    Erik Brynjolfsson, Xiang Hui, and Meng Liu. 2019. Does machine translation affect international trade? evidence from a large digital platform. Management Science, 65(12):5449--5460

    work page 2019

  7. [7]

    In \^e s Carvalho, Ana Ramires, and Montserrat Iglesias. 2023. Attitudes towards machine translation and languages among travelers. Information Technology & Tourism, 25(2):175--204

    work page 2023

  8. [8]

    Chang, Catherine Arnett, Zhuowen Tu, and Benjamin K

    Tyler A. Chang, Catherine Arnett, Zhuowen Tu, and Benjamin K. Bergen. 2024. https://doi.org/10.48550/arXiv.2408.10441 Goldfish: Monolingual Language Models for 350 Languages . arXiv preprint. ArXiv:2408.10441

    arXiv doi:10.48550/arxiv.2408.10441 2024

Show all 43 references
  1. [9]

    Raj Dabre, Chenhui Chu, and Anoop Kunchukuttan. 2020. A survey of multilingual neural machine translation. ACM Computing Surveys (CSUR), 53(5):1--38

    2020

  2. [10]

    Gabriel de Jesus. 2024. http://arxiv.org/abs/2406.07331 Text Information Retrieval in Tetun : A Preliminary Study . arXiv preprint. ArXiv:2406.07331 [cs]

    2024 arXiv

  3. [11]

    Gabriel de Jesus and S \'e rgio Sobral Nunes. 2024 a . https://aclanthology.org/2024.lrec-main.390 Data collection pipeline for low-resource languages: A case study on constructing a tetun text corpus . In Proceedings of the 2024 Joint International Conference on Computational...

    2024

  4. [12]

    Gabriel de Jesus and Sérgio Nunes. 2024 b . https://doi.org/10.48550/arXiv.2406.07299 Exploring Large Language Models for Relevance Judgments in Tetun . arXiv preprint. ArXiv:2406.07299 [cs]

    2024 doi arXiv

  5. [13]

    Gabriel de Jesus and Sérgio Nunes. 2024 c . https://aclanthology.org/2024.sigul-1.22 Labadain-30k+: A Monolingual Tetun Document - Level Audited Dataset . In Proceedings of the 3rd Annual Meeting of the Special Interest Group on Under -resourced Languages @ LREC - COLING 2024 ...

    2024

  6. [14]

    Xinjie Deng and Zhonggen Yu. 2022. https://doi.org/10.3390/su14137598 A Systematic Review of Machine - Translation - Assisted Language Learning for Sustainable Education . Sustainability, 14(13):7598. Number: 13 Publisher: Multidisciplinary Digital Publishing Institute

    2022 doi

  7. [15]

    Dew, Anne M

    Kristin N. Dew, Anne M. Turner, Yong K. Choi, Alyssa Bosold, and Katrin Kirchhoff. 2018. https://doi.org/10.1016/j.jbi.2018.07.018 Development of machine translation technology for assisting health communication: A systematic review . Journal of Biomedical Informatics, 85:56--67

    2018 doi

  8. [16]

    Joanna Dolinska, Shekhar Nayak, and Sumittra Suraratdecha. 2024. https://aclanthology.org/2024.computel-1.13 Akha, Dara -ang, Karen , Khamu , Mlabri and Urak Lawoi ' language minorities' subjective perception of their languages and the outlook for development of digital tools ...

    2024

  9. [17]

    Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783

    2024 arXiv

  10. [18]

    Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, and Armand Joulin. 2020. https://a...

    2020 arXiv

  11. [19]

    James Fiumara, Christopher Cieri, Mark Liberman, Chris Callison-Burch, Jonathan Wright, and Robert Parker. 2022. https://aclanthology.org/2022.nidcp-1.1 The NIEUW Project : Developing Language Resources through Novel Incentives . In Proceedings of the 2nd Workshop on Novel Inc...

    2022

  12. [20]

    General Directorate of Statistics Timor-Leste . 2015. https://inetl-ip.gov.tl/2023/03/09/census-2015-priority-table-population-by-language/ 2015 population and housing census

    2015

  13. [21]

    Zuzana Greksakova. 2018. https://estudogeral.uc.pt/handle/10316/80665 Tetun in Timor - Leste : the role of language contact in its development . doctoralThesis , 00500::Universidade de Coimbra. Accepted: 2018-09-03T09:37:10Z

    2018

  14. [22]

    Barry Haddow, Rachel Bawden, Antonio Valerio Miceli Barone, Jindřich Helcl, and Alexandra Birch. 2022. https://doi.org/10.1162/coli_a_00446 Survey of Low - Resource Machine Translation . Computational Linguistics, 48(3):673--732. Place: Cambridge, MA Publisher: MIT Press

    2022 doi

  15. [23]

    Paula Sofia Herrera-Espejel and Stefan Rach. 2023. https://doi.org/10.2196/50814 The Use of Machine Translation for Outreach and Health Communication in Epidemiology and Public Health : Scoping Review . JMIR Public Health and Surveillance, 9(1):e50814. Company: JMIR Public Hea...

    2023 doi

  16. [24]

    Instituto Nacional de Linguística . 2002. Matadalan Ortográfiku ba Tetun Nasionál . Direcção do Instituto Nacional de Linguística, RDTL , Dili

    2002

  17. [25]

    Seza Do g ru \"o z, and En-Shiun Lee

    Eric Khiu, Hasti Toossi, David Anugraha, Jinyu Liu, Jiaxu Li, Juan Flores, Leandro Roman, A. Seza Do g ru \"o z, and En-Shiun Lee. 2024. https://aclanthology.org/2024.findings-eacl.100 Predicting machine translation performance on low-resource languages: The role of domain sim...

    2024

  18. [26]

    Sneha Kudugunta, Isaac Caswell, Biao Zhang, Xavier Garcia, Derrick Xin, Aditya Kusupati, Romi Stella, Ankur Bapna, and Orhan Firat. 2023. https://proceedings.neurips.cc/paper_files/paper/2023/file/d49042a5d49818711c401d34172f9900-Paper-Datasets_and_Benchmarks.pdf Madlad-400: A...

    2023

  19. [27]

    Moritz Laurer, Wouter van Atteveldt, Andreu Casas, and Kasper Welbers. 2023. https://doi.org/10.48550/arXiv.2312.17543 Building Efficient Universal Classifiers with Natural Language Inference . arXiv preprint. ArXiv:2312.17543 [cs]

    2023 doi arXiv

  20. [28]

    Éric Le Ferrand, Steven Bird, and Laurent Besacier. 2022. https://aclanthology.org/2022.coling-1.376 Fashioning Local Designs from Generic Speech Technologies in an Australian Aboriginal Community . In Proceedings of the 29th International Conference on Computational Linguisti...

    2022

  21. [29]

    Heather Lent, Kelechi Ogueji, Miryam de Lhoneux, Orevaoghene Ahia, and Anders Søgaard. 2022. https://aclanthology.org/2022.lrec-1.691 What a Creole Wants , What a Creole Needs . In Proceedings of the Thirteenth Language Resources and Evaluation Conference , pages 6439--6449, M...

    2022

  22. [30]

    Zoey Liu, Crystal Richardson, Richard Hatcher, and Emily Prud'hommeaux. 2022. https://doi.org/10.18653/v1/2022.acl-long.272 Not always about you: Prioritizing community needs when developing endangered language technology . In Proceedings of the 60th Annual Meeting of the Asso...

    2022 doi

  23. [31]

    Manuel Mager, Elisabeth Mager, Katharina Kann, and Ngoc Thang Vu. 2023. https://doi.org/10.18653/v1/2023.acl-long.268 Ethical Considerations for Machine Translation of Indigenous Languages : Giving a Voice to the Speakers . In Proceedings of the 61st Annual Meeting of the Asso...

    2023 doi

  24. [32]

    Ali Marashian, Enora Rice, Luke Gessler, Alexis Palmer, and Katharina von der Wense. 2025. https://aclanthology.org/2025.coling-main.472/ From Priest to Doctor : Domain Adaptation for Low - Resource Neural Machine Translation . In Proceedings of the 31st International Conferen...

    2025

  25. [33]

    Rapha \"e l Merx, Christine Phillips, and Hanna Suominen. 2024. https://doi.org/10.3233/SHTI240895 Machine translation technology in health: A scoping review . Studies in Health Technology and Informatics, 318:78--83

    2024 doi

  26. [34]

    Hellina Hailu Nigatu, Atnafu Tonja, Benjamin Rosman, Thamar Solorio, and Monojit Choudhury. 2024. https://aclanthology.org/2024.emnlp-main.983 The Zeno 's Paradox of ` Low - Resource ' Languages . In Proceedings of the 2024 Conference on Empirical Methods in Natural Language P...

    2024

  27. [35]

    Kate Paterson. 2023. Machine translation in higher education: Perceptions, policy, and pedagogy. TESOL Journal, 14(2):e690

    2023

  28. [36]

    Surangika Ranathunga, En-Shiun Annie Lee, Marjana Prifti Skenduli, Ravi Shekhar, Mehreen Alam, and Rishemjit Kaur. 2023. https://doi.org/10.1145/3567592 Neural Machine Translation for Low -resource Languages : A Survey . ACM Comput. Surv., 55(11):229:1--229:37

    2023 doi

  29. [37]

    Danielle Saunders. 2022. https://doi.org/10.1613/jair.1.13566 Domain Adaptation and Multi - Domain Adaptation for Neural Machine Translation : A Survey . Journal of Artificial Intelligence Research, 75:351--424

    2022 doi

  30. [38]

    Florian Jaeger

    Job Schepens, Roeland van Hout, and T. Florian Jaeger. 2020. https://doi.org/10.1016/j.cognition.2019.104056 Big data suggest strong constraints of linguistic similarity on adult language learning . Cognition, 194:104056

    2020

  31. [39]

    Roberts, and Lars Yencken

    Hedvig Skirgård, Seán G. Roberts, and Lars Yencken. 2017. https://doi.org/10.1371/journal.pone.0165934 Why are some languages confused for others? Investigating data from the Great Language Game . PLOS ONE, 12(4):e0165934. Publisher: Public Library of Science

    2017 doi

  32. [40]

    The Asia Foundation . 2023. https://asiafoundation.org/publication/digital-youth-in-timor-leste/ D igital Y outh in T imor- L este

    2023

  33. [41]

    Timor-Leste National Institute of Statistics (INETL) . 2022. https://www.laohamutuk.org/DVD/DGS/Cens22/Thematic-Report-on-Migration.pdf Timor- Leste Population and Housing Census 2022 - Thematic Report Migration . Technical report, INETL

    2022

  34. [42]

    Lucas Nunes Vieira, Minako O’Hagan, and Carol O’Sullivan. 2021. https://doi.org/10.1080/1369118X.2020.1776370 Understanding the societal impacts of machine translation: a critical review of the literature on medical and legal use cases . Information, Communication & Society, 2...

    2021

  35. [43]

    Jitao Xu, Josep-Maria Crego, and Jean Senellart. 2020. Boosting neural machine translation with similar translations. In Annual Meeting of the Association for Computational Linguistics, pages 1570--1579. Association for Computational Linguistics

    2020

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.