Pith. sign in

REVIEW 1 major objections 7 minor 43 references

Reviewed by Pith at T0; open to challenge.

T0 means a machine referee read the full paper against a public rubric. The mark states how deep the mechanical check went, never who wrote it. the ladder, T0–T4 →

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

T0 review · glm-5.2

Pivoting through Cyrillic lifts Traditional Mongolian translation

2026-07-08 22:32 UTC pith:37QRABXP

load-bearing objection CoPiT is a solid, well-engineered pipeline for Traditional Mongolian MT with a real linguistic contribution, but the headline comparison to GPT-4.1 is apples-to-oranges and the reference-based evaluation has a structural bias the paper doesn't acknowledge. the 1 major comments →

arxiv 2607.05849 v1 pith:37QRABXP submitted 2026-07-07 cs.CL

CoPiT: Cognitive Pivot Translation for Digraphic Low-Resource Mongolian in the Traditional Script

Burte Bayarsaikhan , Serynn Kim , Buru Chang This is my paper
classification cs.CL
keywords translationscriptcopitmongoliantraditionalcyrillicdatalow-resource
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper proposes CoPiT, a translation pipeline that exploits an internal resource hierarchy within Mongolian itself. Mongolian is digraphic: the Cyrillic script is comparatively well-resourced and phonemically transparent, while the Traditional script is data-scarce and orthographically ambiguous — a single written form can correspond to multiple plausible interpretations. Rather than translating directly from the Traditional script, CoPiT first converts Traditional Mongolian into Cyrillic through a structured multi-step disambiguation process — vowel harmony recovery, Latin-assisted phonological normalization, Cyrillic normalization, and sentence-level self-reflection — and then translates from the disambiguated Cyrillic into the target language. This decomposition isolates script-level ambiguity from semantic transfer, mirroring how fluent Mongolian readers reportedly process the Traditional script by implicitly mapping it to Cyrillic. The authors show that CoPiT consistently outperforms direct translation across multiple backbone models (Qwen-3 4B/30B, Ministral-3 3B/14B, GPT-4.1) and target languages (English, Korean, Russian), with fine-tuned open-source models matching or exceeding GPT-4.1's zero-shot direct-translation performance. The pipeline also generates synthetic parallel data from real Traditional-script sources, enabling both forward and reverse-direction translation in a setting where parallel corpora are virtually nonexistent.

Core claim

The central mechanism is script-level pivoting: routing translation through a better-resourced orthographic representation of the same language to resolve ambiguity before semantic transfer. The Traditional-to-Cyrillic conversion is factorized into linguistically motivated sub-steps — vowel harmony recovery narrows phonological interpretations, Latin-assisted normalization makes implicit phonological distinctions explicit, Cyrillic normalization produces a canonical intermediate form, and self-reflection enforces sentence-level coherence. The ablation shows that self-reflection is the most load-bearing component: removing it causes the largest performance drop (e.g., English COMET falls from

What carries the argument

CoPiT pipeline: morphological segmentation → vowel harmony recovery → Latin-assisted normalization → Cyrillic normalization → sentence reconstruction with self-reflection → Cyrillic-to-target translation. Trained component-wise on 14,125 word-level lexical pairs and 2,061 sentence-level revision pairs, requiring no sentence-level Traditional-to-target parallel data.

Load-bearing premise

The pipeline depends on the Traditional-to-Cyrillic conversion being accurate enough that conversion errors do not propagate into downstream translation, and the evaluation rests on small datasets (1,031 sentences for reference-based evaluation, 12 sentences for human evaluation) with sometimes low inter-annotator agreement. The comparison that open-source models 'match or outperform GPT-4.1' pits fine-tuned open-source models against zero-shot GPT-4.1, an asymmetry in how

What would settle it

If the Traditional-to-Cyrillic conversion step introduced systematic errors that the downstream translation could not recover from, the pivot would degrade rather than improve translation quality. A direct test would measure conversion error rates and correlate them with end-to-end translation quality degradation.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Script-level pivoting could generalize to other digraphic languages where one script is better-resourced than another (e.g., languages with both Latin and non-Latin orthographies, or classical/modern script pairs).
  • The synthetic data generation loop — using the pipeline to create parallel corpora from monolingual Traditional-script sources — offers a self-bootstrapping path for languages where parallel data collection is prohibitively expensive.
  • The finding that self-reflection is the most critical component suggests that global sentence-level coherence, not local disambiguation, is the bottleneck in translating from orthographically ambiguous scripts.
  • Fine-tuned open-source models matching GPT-4.1 suggests that structured linguistic decomposition can compensate for raw model scale in low-resource settings.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

1 major / 7 minor

Summary. The paper proposes CoPiT, a cognitively motivated pipeline for translating Mongolian text written in the Traditional script into English, Korean, and Russian. The core idea is to route translation through the Cyrillic script, which is more phonemically transparent and better-resourced. The Traditional-to-Cyrillic conversion is factorized into linguistically motivated steps: morphological segmentation, vowel harmony recovery, Latin-assisted normalization, Cyrillic normalization, and sentence-level self-reflection. Each component is independently fine-tuned using word-level lexical data (14,125 entries) and sentence-level revision pairs (2,061). Experiments span multiple backbone models (Qwen-3 4B/30B, Ministral-3 3B/14B, GPT-4.1) and three target languages, evaluated under reference-based, reference-free, and human evaluation protocols. The authors also demonstrate that CoPiT-generated synthetic parallel data (8,034 pairs) enables reverse-direction translation and improves forward translation. Datasets and code are released publicly.

Significance. The paper addresses a genuine and well-motivated problem: the digraphic resource imbalance in Mongolian NLP, where the Traditional script is both orthographically ambiguous and severely under-resourced. The cognitive motivation—mirroring how fluent readers map Traditional forms to Cyrillic—is a reasonable framing. The pipeline is linguistically grounded, with components targeting distinct sources of underspecification (vowel harmony, phonological normalization, sentence-level coherence). The ablation study (Table 3, Tables 9–11) is thorough, isolating both individual and pairwise component contributions. The release of a multi-script parallel corpus and all code is a concrete contribution to the community. The synthetic data generation loop and its validation through both forward and reverse translation experiments (Table 4, Figure 6) demonstrate practical utility beyond the inference-time pipeline.

major comments (1)
  1. §4.1, paragraph on Datasets: The reference translations for the reference-based evaluation set (1,031 sentences) are derived from the Cyrillic side of parallel pairs—'English references are translated from the Cyrillic side and subsequently validated by bilingual speakers.' Since CoPiT explicitly routes Traditional→Cyrillic→Target, its outputs are derived from the same Cyrillic representation that generated the references. Direct translation, by contrast, interprets the Traditional script independently and may produce semantically valid translations that diverge from the Cyrillic-grounded reference, particularly for the orthographically ambiguous forms the paper identifies as the core challenge. This creates a systematic structural bias in reference-based evaluation (Table 1) that favors CoPiT by construction. The paper should explicitly acknowledge this bias and clarify that the COMET/B
minor comments (7)
  1. Table 1: The BLEU-3/4 column header is unusual; standard practice reports BLEU-4 (or BLEU). Clarify what BLEU-3/4 means—is it n-gram order 3 and 4 reported separately?
  2. Table 2: The GPT-4.1 row for Russian appears to have a formatting issue where the COMETKiwi value (0.429) runs into the Adeq. column.
  3. §4.2.2: The ablation discussion notes that removing Vowel Harmony Recovery sometimes yields higher COMET (e.g., Qwen-3 4B English: 0.633 without VHR vs. 0.628 with). The paper attributes this to backbone-dependent behavior, but the interaction is not analyzed further. A brief discussion of why VHR can hurt would strengthen the analysis.
  4. Appendix A.2, Table 5: Fluency inter-annotator agreement for English is α=0.263, which is below the conventional threshold for reliable annotation. The paper discusses this but could note more explicitly that fluency conclusions for English should be treated with caution.
  5. §3.2, Morphological Segmentation: The suffix dictionary is mentioned but its size and coverage are not specified. Providing the number of suffixes would help readers assess the generality of this approach.
  6. Figure 2: The fire emoji symbols are unconventional for a system architecture diagram. Consider replacing with standard notation.
  7. References: The citation 'Tumur-Ochir et al.' (in §2.2) is missing a year in the reference list.

Simulated Author's Rebuttal

1 responses · 0 unresolved

The referee raises a valid concern about structural bias in reference-based evaluation: because reference translations are derived from the Cyrillic side of parallel pairs, and CoPiT routes through Cyrillic, CoPiT's outputs may be systematically favored over direct translation outputs that could be semantically valid but diverge from the Cyrillic-grounded reference. We acknowledge this bias and will revise the manuscript accordingly, while also noting that our reference-free evaluation (Table 2) and human evaluation provide independent evidence that is not subject to this concern.

read point-by-point responses

  1. Referee: §4.1, Datasets: The reference translations for the reference-based evaluation set (1,031 sentences) are derived from the Cyrillic side of parallel pairs. Since CoPiT explicitly routes Traditional→Cyrillic→Target, its outputs are derived from the same Cyrillic representation that generated the references. Direct translation interprets the Traditional script independently and may produce semantically valid translations that diverge from the Cyrillic-grounded reference, particularly for orthographically ambiguous forms. This creates a systematic structural bias in reference-based evaluation (Table 1) that favors CoPiT by construction. The paper should explicitly acknowledge this bias and clarify that the COMET/BLEU scores in Table 1 should be interpreted with this caveat.

    Authors: The referee is correct that there is a structural affinity between CoPiT's intermediate representation and the reference translations, since both are derived from the Cyrillic side of the parallel pairs. We acknowledge this bias and will explicitly state it in the revised manuscript, adding a caveat to the description of the reference-based evaluation set in §4.1 and a note in the discussion of Table 1. Specifically, we will add language clarifying that because English references are translated from the Cyrillic side, CoPiT's Cyrillic pivoting may align more closely with reference translations by construction, and that direct translation outputs that are semantically valid but diverge from the Cyrillic-grounded reading would be penalized under reference-based metrics. We will recommend that readers interpret Table 1 results in conjunction with the reference-free evaluation (Table 2), which uses COMETKiwi and human adequacy/fluency ratings and does not rely on Cyrillic-derived references. The reference-free results show consistent improvements for CoPiT across all backbones and target languages, providing evidence that the gains are not solely an artifact of the evaluation bias. That said, we agree the bias should be transparently acknowledged, and the manuscript will be revised accordingly. revision: yes

Circularity Check

0 steps flagged

No significant circularity; evaluation-bias concern is real but is a correctness risk, not a derivation-chain circularity.

full rationale

The paper's derivation chain is: (1) Traditional→Cyrillic conversion trained on word-level lexical pairs (14,125 entries) and sentence-level revision pairs (2,061), (2) Cyrillic→Target translation via LLM. The training data for conversion components is independent of the evaluation test sets (1,031 reference-based, 380 reference-free). The skeptic's headline concern—that English references are translated from the Cyrillic side (Section 4.1: 'English references are translated from the Cyrillic side and subsequently validated by bilingual speakers')—is a legitimate evaluation-bias concern, but it is not circularity in the sense of the derivation reducing to its own inputs. The references are human-created from Cyrillic, not generated by CoPiT itself. CoPiT's outputs are not used to construct the evaluation references. The paper also provides reference-free evaluation (Table 2, COMETKiwi) and human evaluation as independent support that does not depend on Cyrillic-derived references. The synthetic data generation loop (Section 4.2.3) is explicitly acknowledged as an 'upper-bound analysis' and is not presented as a first-principles prediction. No self-citation chain is load-bearing: the cognitive motivation cites an external textbook (Altangerel and Togtokh, 2024), and the self-reflection component cites external work (Wang et al., 2024; Chen et al., 2024). The one point is assigned for the minor structural affinity between the Cyrillic-derived references and the Cyrillic-pivoting pipeline, which, while not circular, could inflate reference-based metrics relative to direct translation—a concern the paper does not fully address but which falls under evaluation validity rather than circularity of derivation.

Axiom & Free-Parameter Ledger

2 free parameters · 4 axioms · 0 invented entities

No new entities, particles, forces, or dimensions are introduced. The pipeline uses existing LLM architectures and linguistic knowledge.

free parameters (2)
  • Suffix dictionary for morphological segmentation = curated, size not specified
    Hand-curated dictionary used to identify suffixes for stem-suffix merging in morphological segmentation (Section 3.2). Not fitted to translation data but is a design choice.
  • LoRA hyperparameters (rank, alpha, dropout) = rank not specified; lr=1e-4; batch=4; grad_accum=2-4; epochs=2-3
    Standard PEFT hyperparameters chosen for fine-tuning each component (Appendix A.3). Minor variations across model scales.
axioms (4)
  • domain assumption Fluent Mongolian readers implicitly map Traditional script to Cyrillic when reading
    Cognitive motivation for the pipeline design, cited to Altangerel and Togtokh (2024), a textbook. Used in Section 1 and 3.1 to justify the pivot architecture.
  • domain assumption Cyrillic Mongolian is sufficiently better-resourced than Traditional to serve as a reliable intermediate representation
    Foundational to the entire pipeline. Supported by the resource imbalance discussion in Section 1 and the experimental results showing Cyrillic-pivoted translation outperforms direct Traditional translation.
  • ad hoc to paper Word-level lexical supervision (14,125 entries) generalizes compositionally to sentence-level disambiguation
    Section 3.4 states components are 'explicitly designed to leverage this word-level lexical supervision and apply it compositionally to longer input sequences.' The self-reflection step partially addresses this, but the assumption that word-level training transfers to sentence-level accuracy is not independently verified.
  • domain assumption BLEU and chrF are adequate metrics for Traditional Mongolian reverse-direction evaluation
    Section 4.2.3 states COMET is unreliable for Traditional Mongolian due to limited script representation in multilingual models, so BLEU/chrF are used instead. This is a pragmatic choice with acknowledged limitations.

pith-pipeline@v1.1.0-glm · 22918 in / 3435 out tokens · 183505 ms · 2026-07-08T22:32:12.338072+00:00 · methodology

0 comments
read the original abstract

Low-resource languages remain challenging for machine translation, and Mongolian is a representative case. As a digraphic language, Mongolian is written in both Cyrillic and Traditional scripts, which exhibit a severe imbalance in data availability. While the Cyrillic script is relatively well-resourced, the Traditional script remains extremely data-scarce and orthographically ambiguous, leading to substantial performance degradation in direct translation. We propose CoPiT, a cognitively motivated pivot-based translation pipeline that exploits this internal resource hierarchy by routing translation through the Cyrillic script. The pipeline explicitly resolves script-induced ambiguity in the Traditional script before translation, enabling more stable and accurate meaning transfer. Across multiple backbone models and target languages, CoPiT consistently outperforms direct translation, achieving substantial absolute BLEU improvements together with consistent 1.5-1.6x COMET gains. These gains allow strong open-source models to match or outperform GPT-4.1 under comparable evaluation settings. Beyond inference-time improvements, CoPiT enables the construction of synthetic parallel data directly from Traditional-script text, mitigating data scarcity in realistic low-resource scenarios. We release a new multi-script parallel dataset covering Mongolian in both scripts alongside English, Korean, and Russian. All datasets and code are publicly available at https://anonymous.4open.science/r/anonymous_project-76C7.

Figures

Figures reproduced from arXiv: 2607.05849 by Burte Bayarsaikhan, Buru Chang, Serynn Kim.

Figure 1
Figure 1. Figure 1: Mongolian digraphia. The same content can be written in Traditional script (left), with multi￾ple surface forms and limited resources, and in better￾resourced Cyrillic script (right). due to limited training data and sparse linguistic resources. This limitation is particularly evident in machine translation (MT), which relies heavily on large-scale parallel corpora that are often unavail￾able for LRLs (Raj… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of CoPiT, a pipeline translating Mongolian from the Traditional script via the Cyrillic script. 2 Related Work 2.1 Low-Resource Language Processing Prior work on low-resource language processing can be broadly categorized into three lines of re￾search. Data-centric methods aim to expand par￾allel data through web mining (Schwenk et al., 2021a,b), back-translation (Sennrich et al., 2016; Edunov et … view at source ↗
Figure 4
Figure 4. Figure 4: Multi-step script disambiguation via vowel [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Sentence-level self-reflection for global se [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: COMET scores under pipeline configurations [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Example of a translation request flagged by [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Full results across BLEU-4, chrF++, and COMET under pipeline configurations with and without [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: COMET scores at increasing CoPiT￾generated synthetic data scales (1K–7K). tween 1K and 3K. Such behavior is characteristic of unstable learning dynamics in extremely low￾resource regimes (Sennrich and Zhang, 2019). Be￾yond this range, performance improvements remain stable across languages despite typological differ￾ences, suggesting that CoPiT-generated data pro￾vides reliable supervision as data scale in… view at source ↗
Figure 10
Figure 10. Figure 10: Prompt used for vowel harmony recovery 18 [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Prompt used for latin normalization 19 [PITH_FULL_IMAGE:figures/full_fig_p019_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Prompt used for Cyrillic normalization 20 [PITH_FULL_IMAGE:figures/full_fig_p020_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Prompt used for self-reflection 21 [PITH_FULL_IMAGE:figures/full_fig_p021_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Prompt used for translation 22 [PITH_FULL_IMAGE:figures/full_fig_p022_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Prompt used for direct translation 23 [PITH_FULL_IMAGE:figures/full_fig_p023_15.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · 7 internal anchors

  1. [1]

    Parallel Corpora for Machine Translation in Low-Resource I ndic Languages: A Comprehensive Review

    Raja, Rahul and Vats, Arpita. Parallel Corpora for Machine Translation in Low-Resource I ndic Languages: A Comprehensive Review. Proceedings of the Eighth Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2025). 2025. doi:10.18653/v1/2025.loresmt-1.12

    work page doi:10.18653/v1/2025.loresmt-1.12 2025

  2. [2]

    2024 , publisher =

    Classical Mongolian: A Textbook for Students, Scholars, and Everyone Interested in Mongolian , author =. 2024 , publisher =

    work page 2024

  3. [3]

    2018 , publisher =

    Үндэсний бичиг VI (Traditional Mongolian Script VI) , author =. 2018 , publisher =

    work page 2018

  4. [4]

    Processing Low-Resource Languages: A Review Of Challenges And Strategies For Inclusive NLP And Sustainable Environment_2025

    Baishya, Diganta and Baruah, Rupam and Bora, Mousoomi and Sarma, Biswajit. Processing Low-Resource Languages: A Review Of Challenges And Strategies For Inclusive NLP And Sustainable Environment_2025. International Journal of Environmental Sciences , year=. doi:10.64252/w55rwj24

    work page doi:10.64252/w55rwj24

  5. [5]

    W iki M atrix: Mining 135 M Parallel Sentences in 1620 Language Pairs from W ikipedia

    Schwenk, Holger and Chaudhary, Vishrav and Sun, Shuo and Gong, Hongyu and Guzm \'a n, Francisco. W iki M atrix: Mining 135 M Parallel Sentences in 1620 Language Pairs from W ikipedia. Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume. 2021. doi:10.18653/v1/2021.eacl-main.115

    work page doi:10.18653/v1/2021.eacl-main.115 2021

  6. [6]

    CCM atrix: Mining Billions of High-Quality Parallel Sentences on the Web

    Schwenk, Holger and Wenzek, Guillaume and Edunov, Sergey and Grave, Edouard and Joulin, Armand and Fan, Angela. CCM atrix: Mining Billions of High-Quality Parallel Sentences on the Web. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume ...

    work page doi:10.18653/v1/2021.acl-long.507 2021

  7. [7]

    Improving Neural Machine Translation Models with Monolingual Data

    Sennrich, Rico and Haddow, Barry and Birch, Alexandra. Improving Neural Machine Translation Models with Monolingual Data. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2016. doi:10.18653/v1/P16-1009

    work page doi:10.18653/v1/p16-1009 2016

  8. [8]

    Understanding Back-Translation at Scale

    Edunov, Sergey and Ott, Myle and Auli, Michael and Grangier, David. Understanding Back-Translation at Scale. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. 2018. doi:10.18653/v1/D18-1045

    work page doi:10.18653/v1/d18-1045 2018

  9. [9]

    Scaling Low-Resource MT via Synthetic Data Generation with LLM s

    de Gibert, Ona and Attieh, Joseph and Vahtola, Teemu and Aulamo, Mikko and Li, Zihao and V. Scaling Low-Resource MT via Synthetic Data Generation with LLM s. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025. doi:10.18653/v1/2025.emnlp-main.1408

    work page doi:10.18653/v1/2025.emnlp-main.1408 2025

  10. [10]

    Multilingual Data Filtering using Synthetic Data from Large Language Models

    Waldendorf, Jonas and Haddow, Barry and Birch, Alexandra and Klimaszewski, Mateusz. Multilingual Data Filtering using Synthetic Data from Large Language Models. Findings of the Association for Computational Linguistics: EMNLP 2025. 2025. doi:10.18653/v1/2025.findings-emnlp.495

    work page doi:10.18653/v1/2025.findings-emnlp.495 2025

  11. [11]

    Unsupervised Cross-lingual Representation Learning at Scale

    Conneau, Alexis and Khandelwal, Kartikay and Goyal, Naman and Chaudhary, Vishrav and Wenzek, Guillaume and Guzm \'a n, Francisco and Grave, Edouard and Ott, Myle and Zettlemoyer, Luke and Stoyanov, Veselin. Unsupervised Cross-lingual Representation Learning at Scale. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. ...

    work page doi:10.18653/v1/2020.acl-main.747 2020

  12. [12]

    m T 5: A Massively Multilingual Pre-trained Text-to-Text Transformer

    Xue, Linting and Constant, Noah and Roberts, Adam and Kale, Mihir and Al-Rfou, Rami and Siddhant, Aditya and Barua, Aditya and Raffel, Colin. m T 5: A Massively Multilingual Pre-trained Text-to-Text Transformer. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2...

    work page doi:10.18653/v1/2021.naacl-main.41 2021

  13. [13]

    T ransli C o: A Contrastive Learning Framework to Address the Script Barrier in Multilingual Pretrained Language Models

    Liu, Yihong and Ma, Chunlan and Ye, Haotian and Schuetze, Hinrich. T ransli C o: A Contrastive Learning Framework to Address the Script Barrier in Multilingual Pretrained Language Models. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024. doi:10.18653/v1/2024.acl-long.136

    work page doi:10.18653/v1/2024.acl-long.136 2024

  14. [14]

    Proceedings of the 5th Workshop on Multilingual Representation Learning (MRL 2025) , pages=

    Exploring the role of transliteration in in-context learning for low-resource languages written in non-latin scripts , author=. Proceedings of the 5th Workshop on Multilingual Representation Learning (MRL 2025) , pages=

    work page 2025

  15. [15]

    and Carbonell, Jaime

    Chaudhary, Aditi and Zhou, Chunting and Levin, Lori and Neubig, Graham and Mortensen, David R. and Carbonell, Jaime. Adapting Word Embeddings to New Languages with Morphological and Phonological Subword Representations. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. 2018. doi:10.18653/v1/D18-1366

    work page doi:10.18653/v1/d18-1366 2018

  16. [16]

    Low-resource neural machine translation with morphological modeling

    Low-resource neural machine translation with morphological modeling , author=. arXiv preprint arXiv:2404.02392 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  17. [17]

    International Conference on Intelligent Multilingual Information Processing , pages=

    MM-Eval: A Hierarchical Benchmark for Modern Mongolian Evaluation in LLMs , author=. International Conference on Intelligent Multilingual Information Processing , pages=. 2024 , organization=

    work page 2024

  18. [18]

    Evaluating Large Language Models in Mongolian , author =

    work page

  19. [19]

    2024 International Joint Conference on Neural Networks (IJCNN) , pages=

    Pre-training Language Model for Mongolian with Agglutinative Linguistic Knowledge Injection , author=. 2024 International Joint Conference on Neural Networks (IJCNN) , pages=. 2024 , organization=

    work page 2024

  20. [20]

    Incorporating Inner-word and Out-word Features for M ongolian Morphological Segmentation

    Liu, Na and Su, Xiangdong and Zhang, Haoran and Gao, Guanglai and Bao, Feilong. Incorporating Inner-word and Out-word Features for M ongolian Morphological Segmentation. Proceedings of the 28th International Conference on Computational Linguistics. 2020. doi:10.18653/v1/2020.coling-main.408

    work page doi:10.18653/v1/2020.coling-main.408 2020

  21. [21]

    Dependent syntactic analysis of Mongolian based on semi-supervised self-training , year=

    Zhang, Zhonghao and Ma, Jiajia and Liu, Na and Wu, Nier and Ji, Yatu and Liu, Guiping , booktitle=. Dependent syntactic analysis of Mongolian based on semi-supervised self-training , year=

    work page

  22. [22]

    The Study of Comparison and Conversion about Traditional Mongolian and Cyrillic Mongolian , year=

    Li, Hao and Sarina, Bao , booktitle=. The Study of Comparison and Conversion about Traditional Mongolian and Cyrillic Mongolian , year=

    work page

  23. [23]

    2023 IEEE 9th International Conference on Cloud Computing and Intelligent Systems (CCIS) , pages=

    Traditional Mongolian-to-Cyrillic Mongolian Conversion Method Based on the Combination of Rules and Transformer , author=. 2023 IEEE 9th International Conference on Cloud Computing and Intelligent Systems (CCIS) , pages=. 2023 , organization=

    work page 2023

  24. [24]

    Journal of Institute of Mathematics and Digital Technology , author=

    Cyrillic Mongolian-to-traditional Mongolian conversion method based on the transformer , volume=. Journal of Institute of Mathematics and Digital Technology , author=. 2024 , month=. doi:10.5564/jimdt.v6i1.3599 , abstractNote=

    work page doi:10.5564/jimdt.v6i1.3599 2024

  25. [25]

    International Conference on Neural Information Processing , pages=

    A Deep Investigation of RNN and Self-attention for the Cyrillic-Traditional Mongolian Bidirectional Conversion , author=. International Conference on Neural Information Processing , pages=. 2022 , organization=

    work page 2022

  26. [26]

    Journal of Institute of Mathematics and Digital Technology , volume=

    Joint NMT models for text conversion between traditional Mongolian script and cyrillic Mongolian: a comparative study , author=. Journal of Institute of Mathematics and Digital Technology , volume=

    work page

  27. [27]

    , title =

    Janhunen, Juha A. , title =. 2012 , pages =

    work page 2012

  28. [28]

    Qwen3 Technical Report

    Qwen3 technical report , author=. arXiv preprint arXiv:2505.09388 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  29. [29]

    2025 , howpublished =

    Introducing Mistral 3 , author =. 2025 , howpublished =

    work page 2025

  30. [30]

    2024 , howpublished =

    GPT-4.1 , author =. 2024 , howpublished =

    work page 2024

  31. [31]

    GPT-4 Technical Report

    Gpt-4 technical report , author=. arXiv preprint arXiv:2303.08774 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  32. [32]

    Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities

    Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities , author=. arXiv preprint arXiv:2507.06261 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  33. [33]

    Proceedings of the 40th annual meeting of the Association for Computational Linguistics , pages=

    Bleu: a method for automatic evaluation of machine translation , author=. Proceedings of the 40th annual meeting of the Association for Computational Linguistics , pages=

    work page

  34. [34]

    Proceedings of the tenth workshop on statistical machine translation , pages=

    chrF: character n-gram F-score for automatic MT evaluation , author=. Proceedings of the tenth workshop on statistical machine translation , pages=

    work page

  35. [35]

    Proceedings of the second conference on machine translation , pages=

    chrF++: words helping character n-grams , author=. Proceedings of the second conference on machine translation , pages=

    work page

  36. [36]

    Proceedings of the Seventh Conference on Machine Translation (WMT) , pages=

    COMET-22: Unbabel-IST 2022 submission for the metrics shared task , author=. Proceedings of the Seventh Conference on Machine Translation (WMT) , pages=

    work page 2022

  37. [37]

    Proceedings of the Eighth Conference on Machine Translation , pages=

    Scaling up cometkiwi: Unbabel-ist 2023 submission for the quality estimation shared task , author=. Proceedings of the Eighth Conference on Machine Translation , pages=

    work page 2023

  38. [38]

    Edward J Hu and Yelong Shen and Phillip Wallis and Zeyuan Allen-Zhu and Yuanzhi Li and Shean Wang and Lu Wang and Weizhu Chen , booktitle=. Lo. 2022 , url=

    work page 2022

  39. [39]

    TasTe: Teaching Large Language Models to Translate through Self-Reflection

    Taste: Teaching large language models to translate through self-reflection , author=. arXiv preprint arXiv:2406.08434 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  40. [40]

    DUAL-REFLECT: Enhancing Large Language Models for Reflective Translation through Dual Learning Feedback Mechanisms

    DUAL-REFLECT: Enhancing large language models for reflective translation through dual learning feedback mechanisms , author=. arXiv preprint arXiv:2406.07232 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  41. [41]

    International conference on machine learning , pages=

    Similarity of neural network representations revisited , author=. International conference on machine learning , pages=. 2019 , organization=

    work page 2019

  42. [42]

    Revisiting Low-Resource Neural Machine Translation: A Case Study

    Revisiting low-resource neural machine translation: A case study , author=. arXiv preprint arXiv:1905.11901 , year=

    work page internal anchor Pith review Pith/arXiv arXiv 1905

  43. [43]

    2019 , edition =

    Content Analysis: An Introduction to Its Methodology , author =. 2019 , edition =. doi:10.4135/9781071878781 , url =

    work page doi:10.4135/9781071878781 2019