Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Sinhala Transliteration: A Comparative Analysis Between Rule-based and Seq2Seq Approaches

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Fine-tuning a multilingual translation model beats hand-written rules for Singlish-to-Sinhala transliteration.

desk verdict Useful benchmark for Sinhala transliteration, but the BLEU gap needs an overlap check before the headline numbers are trustworthy. read the letter →

arxiv 2501.00529 v1 pith:QWMQW3TN submitted 2024-12-31 cs.CL

Yomal De Mel , Kasun Wickramasinghe , Nisansa de Silva , Surangika Ranathunga This is my paper · ORCID
classification cs.CL
keywords SinhalatransliterationSinglishrule-basedsequence-to-sequenceM2M100low-resourceNLPreverseBLEU
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 tries to establish that Romanized Sinhala (Singlish) to Sinhala script is better handled by a fine-tuned sequence-to-sequence model than by hand-authored transliteration rules. The authors compare a rule-based character-mapping system with a fine-tuned M2M100 model, trained on just 9,000 parallel examples, and evaluate both on two shared-task test sets. The neural model achieves average BLEU of 0.5268 on the general test set and 0.4384 on the ad-hoc set, while the rule-based system scores 0.0177 and 0.0163. The authors argue this matters because Singlish typing is highly variable, code-mixed, and often drops vowels, making fixed rules brittle. If correct, the result offers a practical recipe: a small fine-tune of an existing multilingual translation model can serve as a reverse transliterator for a low-resource language without designing new linguistic rules.

What carries the argument

The deep-learning system is a fine-tuned M2M100 encoder-decoder transformer, a multilingual sequence-to-sequence model pretrained for translation; the authors reuse its English tokenizer by treating Romanized Sinhala as 'English' and Sinhala as the target language. Fine-tuning is done with a learning rate of 2e-5 for three epochs on 9,000 training pairs. The rule-based baseline is a longest-match lookup table that extends the mapping rules of an existing Sinhala transliteration system with additional two- and three-character sequences; it greedily consumes the longest matched substring at each position. The M2M100 model is the machinery that carries the accuracy claim, while the table carries the baseline comparison.

What would settle it

Re-run the evaluation after checking every training and validation sentence against the two shared-task test sets and removing any near-duplicate overlap; if the neural BLEU drops substantially from 0.5268 and 0.4384 when overlaps are excluded, the claimed advantage is partly memorization. A second check is to test the same fine-tuned model on a freshly collected, non-overlapping Singlish corpus, which would show whether the model generalises beyond the benchmark's distribution.

Watch

Extended reading notes

Core claim

The paper's central claim is that the reverse transliteration task for Singlish to Sinhala should be treated as a machine translation problem rather than a lookup over fixed character mappings. Fine-tuning the 418M-parameter M2M100 model on 9,000 parallel examples yields an average BLEU of 0.5268 on test set 01 and 0.4384 on test set 02, compared with 0.0177 and 0.0163 for a rule-based system that extends an existing mapping table with additional two- and three-character rules. The neural model also reports lower word error rate and character error rate on both test sets, and handles vowel-dropped and ad-hoc spellings that the rule-based approach gets wrong. The paper concludes that the deep-learning approach is more robust to language variability, while the rule-based approach remains faster on CPU and unsuitable for real-time use without GPU acceleration.

Load-bearing premise

The comparison assumes the shared-task test sets are an external benchmark that does not overlap with the 9,000 training and 1,000 validation pairs built from Dakshina and Swa-Bhasha; no deduplication check is reported.

Editorial extensions

If this is right

  • With only 9,000 parallel training examples, a fine-tuned multilingual translation model can outperform a hand-tuned transliteration table by a large margin on held-out test sets.
  • Because the sequence-to-sequence model uses context, it can resolve ambiguous and vowel-dropped spellings that fixed character mappings cannot.
  • On GPU the neural approach is real-time (about 65 tokens per second in fp16), while the rule-based approach is much faster on CPU but far less accurate.
  • The same training recipe should transfer to other low-resource languages with Romanized typing conventions, since it only needs a parallel corpus rather than language-specific rules.
  • The lower BLEU on test set 02 (0.4384 versus 0.5268) shows that ad-hoc typing remains harder, so accuracy on extreme spellings is still an open issue.

Reading between the lines

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

  • Editorial inference: even a substantial training/test overlap would need to be very large to erase the reported gap, but a deduplicated benchmark is needed to know the true effect size.
  • Editorial inference: the paper's choice to encode Romanized Sinhala with the English language token may under-use M2M100's multilingual capacity; a dedicated Romanized-Sinhala token could improve handling of Sinhala-specific phonetics.
  • Editorial inference: a hybrid pipeline that applies rule-based lookup to clear tokens and the neural model only to ambiguous ones could preserve most of the accuracy gain while reducing GPU dependence.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper compares two approaches for transliterating Romanized Sinhala (Singlish) into Sinhala script: a rule-based transliteration system and a fine-tuned multilingual sequence-to-sequence model (M2M100 418M). The rule-based method uses a hand-authored character mapping table; the deep learning method is fine-tuned on 9,000 parallel training pairs constructed from the Dakshina and Swa-Bhasha corpora with additional augmentation that simulates ad-hoc typing patterns. Both methods are evaluated on the IndoNLP 2025 shared-task test sets (10,000 general and 5,000 ad-hoc entries) using BLEU, WER, and CER. The reported results show a large advantage for the deep learning approach (BLEU 0.5268 and 0.4384 on the two test sets versus 0.0177 and 0.0163 for the rule-based system), while the rule-based system is far faster. The paper also discusses limitations and efficiency trade-offs, and provides a public code repository.

Significance. If the comparison is valid, this is a useful empirical data point for low-resource transliteration: it shows that fine-tuning a large multilingual seq2seq model on a modest parallel corpus can substantially outperform hand-written rules on noisy, ad-hoc Romanized input, and the use of two external shared-task test sets makes the evaluation more credible than a self-constructed test would be. The code release supports reproducibility. The significance is tempered, however, by the absence of any train/test overlap analysis, the lack of variance or significance estimates, and an underspecified rule-based baseline; these issues do not necessarily overturn the qualitative conclusion, but they prevent the reader from trusting the specific numbers and the strength of the claim until they are addressed.

major comments (3)
  1. [Section 4.1 and Section 5, Table 5] The central comparison assumes that the IndoNLP 2025 shared-task test sets are genuinely held out from the 9,000 training pairs built from Dakshina and Swa-Bhasha, but the paper reports no deduplication or overlap analysis. This is load-bearing because Swa-Bhasha is a public corpus of Romanized Sinhala containing ad-hoc forms, and the augmentation procedure in Table 3 explicitly generates vowel-dropped and shortened variants (e.g., 'kawd' from 'kawada', 'oy' from 'oya') that are exactly the kind of patterns described for Test Set 02. If even a small fraction of the 15,000 test sentences closely resembles training examples, the reported BLEU gap in Table 5 could be inflated by memorization rather than generalization. The authors should add an explicit overlap check (exact string matching as well as normalized matching after vowel removal and other augmentation-style transformations) between the training/validation data and both test sets, and report the number of overlapping or near-duplicate entries and any actions taken.
  2. [Section 5, Table 5] The evaluation reports a single run of the fine-tuned model and a single deterministic rule-based system, without error bars, multiple seeds, or significance tests. While the BLEU gap is large enough that sampling noise is unlikely to reverse the qualitative ordering, the specific values in Table 5 and the claim that the deep learning approach is 'more robust' (Section 5) need supporting evidence such as standard deviations across multiple runs or bootstrap confidence intervals for the metrics. Without this, the reader cannot assess the stability of the reported averages.
  3. [Section 3.1 and Table 1] The rule-based baseline is described as an extension of Tennage et al. (2018) with 'a few additions,' but the paper does not specify which rules were added, provide the complete mapping table, or list the exact version of the rule set used. Since the paper's core claim is a comparative evaluation, the baseline must be fully specified or unambiguously pinned to a released configuration (e.g., a tagged version in the GitHub repository) so that readers can reproduce or assess the strength of the rule-based system. As written, it is unclear whether this is a reasonable state-of-the-art rule-based baseline or a minimal character-level matcher.
minor comments (6)
  1. [Footnote 1] The use of 'Singlish' to mean Romanized Sinhala is confusing because the same term is commonly used for Singapore Colloquial English; the paper acknowledges this but should use a less ambiguous term or consistently write 'Singlish (Romanized Sinhala)' throughout.
  2. [Table 2] Several rows in the qualitative comparison table contain garbled or inconsistently ordered Sinhala text (e.g., 'ෙම් ෙද්'), which could be a PDF rendering issue. If the text is corrupted, the table should be regenerated; if not, it undermines the qualitative comparison of the pre-trained models.
  3. [Section 4.1] The sentence 'The original datasets were not well structured' is ambiguous: it is unclear whether it refers to the Dakshina and Swa-Bhasha training sources, the shared-task test sets, or both. Please clarify and describe the conversion to CSV more precisely.
  4. [Algorithm 1] The pseudocode does not state what happens to Latin characters that do not match any rule. The text says they are 'added as is,' but it is unclear whether these characters are inserted into the Sinhala output as Latin characters and whether this is consistent with the evaluation setting.
  5. [Section 2.1 and Section 3.1] The paper attributes the first Sinhala-to-English transliteration system and the rule-based system used as a baseline to Tennage et al. (2018), but that reference describes a Tamil-to-Sinhala neural machine translation system using transliteration and byte-pair encoding, not a Sinhala-to-English transliteration tool. Please correct or rephrase this citation.
  6. [Table 7] The speed comparison reports TPS values but does not specify the hardware for the CPU row (the GPU rows are labeled, but the CPU model is not), nor the exact input used for the measurement. Please add these details so the efficiency comparison is reproducible.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central comparison is an empirical evaluation on external shared-task test sets with independently sourced training data.

full rationale

The paper's central claim is that a fine-tuned M2M100 model outperforms a rule-based transliterator on the IndoNLP 2025 shared-task test sets. The evaluation is external: the test sets are provided by the shared task organizers, and the fine-tuning data are built from Dakshina and Swa-Bhasha, which are public corpora distinct from the test sets as described. The rule-based method is defined by a fixed transliteration table and a longest-match algorithm, not by any parameter fitted to the reported BLEU/WER/CER values. The neural model's weights are trained on the 9k training split and validated on the 1k validation split, and the reported numbers are averages over the held-out test sets. There is no equation in which a reported result is defined in terms of the input, and no load-bearing self-citation chain: citations to prior work by the same research group are contextual (e.g., model selection rationale) rather than the source of the claimed result. The absence of an explicit overlap check between training and test data is a legitimate correctness or data-hygiene concern, but it is not circularity under the defined criteria, because the test sets are still presented as external and the paper does not fit any parameter to them. Therefore the paper is self-contained with respect to circularity concerns, and the appropriate score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new entities or fitted constants. Its central claim depends on two external data assumptions (training corpus quality and test set fairness) plus a modeling-choice assumption about M2M100's language codes.

assumptions (3)
  • domain assumption The Dakshina and Swa-Bhasha datasets contain correct and representative Romanized-Sinhala to Sinhala pairs.
    The training and validation data are built from these sources without any stated quality filtering or error analysis beyond the authors' augmentation.
  • domain assumption The IndoNLP 2025 shared-task test sets are accurate, unbiased samples of real Singlish usage.
    The evaluation claim rests entirely on these held-out sets; no independent validation is reported.
  • domain assumption M2M100's English language code and tokenizer can adequately represent Romanized Sinhala input.
    The authors choose the 'en' code for Singlish, which is a pragmatic but unvalidated modeling choice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Sinhala Transliteration: A Comparative Analysis Between Rule-based and Seq2Seq Approaches." pith.science (2026). https://pith.science/paper/QWMQW3TN

@misc{pith2026250100529,
  author       = {Pith},
  title        = {Pith review of: Sinhala Transliteration: A Comparative Analysis Between Rule-based and Seq2Seq Approaches},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QWMQW3TN}},
  note         = {Machine review of arXiv:2501.00529}
}
read the original abstract

Due to reasons of convenience and lack of tech literacy, transliteration (i.e., Romanizing native scripts instead of using localization tools) is eminently prevalent in the context of low-resource languages such as Sinhala, which have their own writing script. In this study, our focus is on Romanized Sinhala transliteration. We propose two methods to address this problem: Our baseline is a rule-based method, which is then compared against our second method where we approach the transliteration problem as a sequence-to-sequence task akin to the established Neural Machine Translation (NMT) task. For the latter, we propose a Transformer-based Encode-Decoder solution. We witnessed that the Transformer-based method could grab many ad-hoc patterns within the Romanized scripts compared to the rule-based method. The code base associated with this paper is available on GitHub - https://github.com/kasunw22/Sinhala-Transliterator/

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. IndoNLP 2025: Shared Task on Real-Time Reverse Transliteration for Romanized Indo-Aryan languages

    cs.CL 2025-01 conditional novelty 4.0 of 10

    A shared task benchmark for Romanized to native script transliteration across Sinhala, Hindi, and Malayalam, with the organizers' team winning Sinhala.

Reference graph

Works this paper leans on

41 extracted references · 28 canonical work pages · cited by 1 Pith paper

  1. [1]

    Maneesha U Athukorala and Deshan K Sumanathilaka. 2024. Swa bhasha: Message-based singlish to sinhala transliteration. arXiv preprint arXiv:2404.13350

    work page Pith review arXiv 2024

  2. [2]

    Marta R Costa-juss \`a , James Cross, Onur C elebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, et al. 2022. No language left behind: Scaling human-centered machine translation. arXiv preprint arXiv:2207.04672

    arXiv 2022

  3. [3]

    Melonie de Almeida, Chamodi Samarawickrama, Nisansa de Silva, Gathika Ratnayaka, and Amal Shehan Perera. 2020. Legal Party Extraction from Legal Opinion Text with Sequence to Sequence Learning . In 2020 20th International Conference on Advances in ICT for Emerging Regions (ICTer), pages 143--148. IEEE

    work page 2020

  4. [4]

    Nisansa De Silva. 2019. Survey on publicly available sinhala natural language processing tools and research. arXiv preprint arXiv:1906.02358

    arXiv 2019

  5. [5]

    Thomas Deselaers, Sa s a Hasan, Oliver Bender, and Hermann Ney. 2009. https://aclanthology.org/W09-0438 A deep learning approach to machine transliteration . In Proceedings of the Fourth Workshop on Statistical Machine Translation, pages 233--241, Athens, Greece. Association for Computational Linguistics

    work page 2009

  6. [6]

    Ahmed El-Kishky, Vishrav Chaudhary, Francisco Guzman, and Philipp Koehn. 2019. A massive collection of cross-lingual web-document pairs. arXiv preprint arXiv:1911.06154

    arXiv 2019

  7. [7]

    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. Beyond english-centric multilingual machine translation. arXiv preprint

    work page 2020

  8. [8]

    Kamaljeet Kaur and Parminder Singh. 2014. Review of machine transliteration techniques. International Journal of Computer Applications, 107(20)

    work page 2014

Show all 41 references
  1. [9]

    Soumyadeep Kundu, Sayantan Paul, and Santanu Pal. 2018. https://doi.org/10.18653/v1/W18-2411 A deep learning based approach to transliteration . In Proceedings of the Seventh Named Entities Workshop, pages 79--83, Melbourne, Australia. Association for Computational Linguistics

    2018 doi

  2. [10]

    En-Shiun Lee, Sarubi Thillainathan, Shravan Nayak, Surangika Ranathunga, David Adelani, Ruisi Su, and Arya D McCarthy. 2022. Pre-trained multilingual sequence-to-sequence models: A hope for low-resource language translation? In Findings of the Association for Computational Lin...

    2022

  3. [11]

    M Lewis. 2019. Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. arXiv preprint arXiv:1910.13461

    2019 arXiv

  4. [12]

    Y Liu. 2020. Multilingual denoising pre-training for neural machine translation. arXiv preprint arXiv:2001.08210

    2020 arXiv

  5. [13]

    WMP Liwera and L Ranathunga. 2020. Combination of trigram and rule-based model for singlish to sinhala transliteration by focusing social media text. In 2020 From Innovation to Impact (FITI), volume 1, pages 1--5. IEEE

    2020

  6. [14]

    Chamila Liyanage, Randil Pushpananda, Dulip Lakmal Herath, and Ruvan Weerasinghe. 2012. A computational grammar of sinhala. In Computational Linguistics and Intelligent Text Processing: 13th International Conference, CICLing 2012, New Delhi, India, March 11-17, 2012, Proceedin...

    2012

  7. [15]

    Mohammad Mahdi Mahsuli and Reza Safabakhsh. 2017. https://doi.org/10.1109/IranianCEE.2017.7985375 English to persian transliteration using attention-based approach in deep learning . In 2017 Iranian Conference on Electrical Engineering (ICEE), pages 174--178

    2017

  8. [16]

    Molly Moran and Constantine Lignos. 2020. https://aclanthology.org/2020.loresmt-1.11 Effective architectures for low resource multilingual named entity transliteration . In Proceedings of the 3rd Workshop on Technologies for MT of Low Resource Languages, pages 79--86, Suzhou, ...

    2020

  9. [17]

    Kashyapa Niyarepola, Dineth Athapaththu, Savindu Ekanayake, and Surangika Ranathunga. 2022. Math word problem generation with multilingual language models. In Proceedings of the 15th International Conference on Natural Language Generation, pages 144--155

    2022

  10. [18]

    Ernst Pulgram. 1951. Phoneme and grapheme: A parallel . Word, 7(1):15--20

    1951

  11. [19]

    Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research, 21(140):1--67

    2020

  12. [20]

    Surangika Ranathunga, Nisansa De Silva, Velayuthan Menan, Aloka Fernando, and Charitha Rathnayake. 2024 a . https://aclanthology.org/2024.eacl-long.52 Quality does matter: A detailed look at the quality and utility of web-mined parallel corpora . In Proceedings of the 18th Con...

    2024

  13. [21]

    Surangika Ranathunga, En-Shiun Annie Lee, Marjana Prifti Skenduli, Ravi Shekhar, Mehreen Alam, and Rishemjit Kaur. 2023. Neural machine translation for low-resource languages: A survey. ACM Computing Surveys, 55(11):1--37

    2023

  14. [22]

    Surangika Ranathunga, Rumesh Sirithunga, Himashi Rathnayake, Lahiru De Silva, Thamindu Aluthwala, Saman Peramuna, and Ravi Shekhar. 2024 b . Sitse: Sinhala text simplification dataset and evaluation. arXiv preprint arXiv:2412.01293

    2024 arXiv

  15. [23]

    Himashi Rathnayake, Janani Sumanapala, Raveesha Rukshani, and Surangika Ranathunga. 2022. Adapter-based fine-tuning of pre-trained multilingual language models for code-mixed and code-switched text classification. Knowledge and Information Systems, 64(7):1937--1966

    2022

  16. [24]

    Mielke, Cibu Johny, I s in Demir s ahin, and Keith Hall

    Brian Roark, Lawrence Wolf-Sonkin, Christo Kirov, Sabrina J. Mielke, Cibu Johny, I s in Demir s ahin, and Keith Hall. 2020. https://www.aclweb.org/anthology/2020.lrec-1.294 Processing South Asian languages written in the Latin script: the Dakshina dataset . In Proceedings of T...

    2020

  17. [25]

    Mihaela Rosca and Thomas Breuel. 2016. Sequence-to-sequence neural network models for transliteration. arXiv preprint arXiv:1610.09565

    2016 arXiv

  18. [26]

    Marianne Santaholma. 2007. https://aclanthology.org/W07-2438 Grammar sharing techniques for rule-based multilingual NLP systems . In Proceedings of the 16th Nordic Conference of Computational Linguistics ( NODALIDA 2007) , pages 253--260, Tartu, Estonia. University of Tartu, Estonia

    2007

  19. [27]

    Holger Schwenk, Guillaume Wenzek, Sergey Edunov, Edouard Grave, Armand Joulin, and Angela Fan. 2021. https://doi.org/10.18653/v1/2021.acl-long.507 CCM atrix: Mining billions of high-quality parallel sentences on the web . In Proceedings of the 59th Annual Meeting of the Associ...

    2021 doi

  20. [28]

    Yan Shao and Joakim Nivre. 2016. https://doi.org/10.18653/v1/W16-2710 Applying neural networks to E nglish- C hinese named entity transliteration . In Proceedings of the Sixth Named Entity Workshop, pages 73--77, Berlin, Germany. Association for Computational Linguistics

    2016 doi

  21. [29]

    Deshan Sumanathilaka, Nicholas Micallef, and Ruvan Weerasinghe. 2024. Swa-bhasha dataset: Romanized sinhala to sinhala adhoc transliteration corpus. In 2024 4th International Conference on Advanced Research in Computing (ICARC), pages 189--194. IEEE

    2024

  22. [30]

    TGDK Sumanathilaka. 2023. Romanized sinhala to sinhala reverse transliteration using a hybrid approach. Ph.D. thesis

    2023

  23. [31]

    TGDK Sumanathilaka, Ruvan Weerasinghe, and YHPP Priyadarshana. 2023. Swa-bhasha: Romanized sinhala to sinhala reverse transliteration using a hybrid approach. In 2023 3rd International Conference on Advanced Research in Computing (ICARC), pages 136--141. IEEE

    2023

  24. [32]

    George Tambouratzis. 2021. https://doi.org/10.18653/v1/2021.eacl-main.158 Alignment verification to improve NMT translation towards highly inflectional languages with limited resources . In Proceedings of the 16th Conference of the European Chapter of the Association for Compu...

    2021 doi

  25. [33]

    Pasindu Tennage, Achini Herath, Malith Thilakarathne, Prabath Sandaruwan, and Surangika Ranathunga. 2018. Transliteration and byte pair encoding to improve tamil to sinhala neural machine translation. In 2018 Moratuwa Engineering Research Conference (MERCon), pages 390--395. IEEE

    2018

  26. [34]

    Sarubi Thillainathan, Surangika Ranathunga, and Sanath Jayasena. 2021. Fine-tuning self-supervised multilingual sequence-to-sequence models for extremely low-resource nmt. In 2021 Moratuwa Engineering Research Conference (MERCon), pages 432--437. IEEE

    2021

  27. [35]

    Pasindu Udawatta, Indunil Udayangana, Chathulanka Gamage, Ravi Shekhar, and Surangika Ranathunga. 2024. Use of prompt-based learning for code-mixed and code-switched text classification. World Wide Web, 27(5):63

    2024

  28. [36]

    A Vaswani. 2017. Attention is all you need. Advances in Neural Information Processing Systems

    2017

  29. [37]

    Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, and Colin Raffel. 2021. https://doi.org/10.18653/v1/2021.naacl-main.41 m T 5: A massively multilingual pre-trained text-to-text transformer . In Proceedings of the 2021 Conferenc...

    2021 doi

  30. [38]

    Biao Zhang, Philip Williams, Ivan Titov, and Rico Sennrich. 2020. https://doi.org/10.18653/v1/2020.acl-main.148 Improving massively multilingual neural machine translation and zero-shot translation . In Proceedings of the 58th Annual Meeting of the Association for Computationa...

    2020 doi

  31. [39]

    Reihaneh Zohrabi, Mostafa Masumi, Omid Ghahroodi, Parham AbedAzad, Hamid Beigy, Mohammad Hossein Rohban, and Ehsaneddin Asgari. 2023. https://doi.org/10.18653/v1/2023.ijcnlp-short.20 Borderless A zerbaijani processing: Linguistic resources and a transformer-based approach for ...

    2023 doi

  32. [40]

    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...

  33. [41]

    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 gl...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.