Pith. sign in

REVIEW 5 major objections 5 minor 52 references

Unveiling Factors for Enhanced POS Tagging: A Study of Low-Resource Medieval Romance Languages

T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read Fine-tuning open LLMs beats prompting for tagging medieval Romance texts, and pooling related languages helps most on the scarcest corpus.

desk verdict Useful new datasets and a broad benchmark, but the cross-lingual transfer claim conflates pooling with transfer, and single-run numbers make the conclusions provisional. read the letter →

arxiv 2506.17715 v1 pith:RVAIF7DZ submitted 2025-06-21 cs.CL cs.LG

Matthias Schöffel , Esteban Garces Arias , Marinus Wiedner , Paula Ruppert , Meimingwei Li , Christian Heumann , Matthias A{ss}enmacher This is my paper · ORCID
classification cs.CLcs.LG
keywords part-of-speechtaggingmedievalRomancelanguageslow-resourceNLPcross-lingualtransferlearningfine-tuningvspromptinghistoricaltextprocessingOccitanopen-sourceLLMs
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 asks what actually determines part-of-speech tagging accuracy when large language models are applied to Medieval Occitan, Medieval French, and Medieval Spanish, varieties with erratic spelling, dialectal mixture, and very little annotated data. It reports that fine-tuning an open instruction-tuned model on annotated data consistently outperforms prompting with examples, and that training on several medieval Romance languages together improves accuracy on average by +0.86 percentage points over single-dataset training. The largest gain, +5.69 points, lands on the Harley 7403 corpus, the lowest-resource dataset in the study. The paper also argues that pre-training language coverage and architecture matter more than parameter count, since an 8-billion-parameter model with Romance training beats 14-billion-parameter models. A careful reader would care because these findings give concrete, evidence-based choices for anyone building NLP tools for historical languages.

What carries the argument

The central mechanism is cross-lingual transfer learning through joint fine-tuning: each model is fine-tuned on 80% of all seven datasets combined, then evaluated on the held-out 20% of one target dataset, and compared against single-dataset fine-tuning under the same split. This pooled-training protocol is what lets related Romance varieties share evidence, and is the object that produces the +0.86 average gain and the +5.69-point Harley improvement. A second mechanism is the prompting evaluation on Gemma2-9B with four decoding strategies (greedy, temperature 0.3, temperature 0.9, and top-p 0.95), which establishes the language-specific decoding recommendations. The load-bearing comparability comes from keeping the fine-tuning split and evaluation metrics identical across all seven models and datasets.

What would settle it

Re-annotate a random sample of each of the seven corpora with two independent expert annotators, compute inter-annotator agreement, and re-run the main fine-tuning and transfer comparisons on corrected labels; if the +5.69-point Harley transfer gain or the fine-tuning-vs-prompting gaps shrink below noise, the paper's central conclusions fail. Alternatively, a domain-matched pooled-training experiment on an independent medieval language pair would test whether the average +0.86 transfer gain replicates.

Watch

Extended reading notes

Core claim

The central claim is that for POS tagging of low-resource medieval Romance languages, fine-tuning is more reliable than prompting, and cross-lingual transfer learning brings a modest but real average gain of +0.86 percentage points over single-dataset fine-tuning. The transfer benefit is sharply heterogeneous: the lowest-resource Occitan corpus, Harley, improves by +5.69 points, Chauliac by +2.61, while Lapidaire drops by -3.79 points, a negative-transfer pattern the paper attributes to limited shared vocabulary. On prompting, few-shot consistently beats zero-shot by an average of +0.0120 accuracy, and the best decoding strategy depends on the language variety: temperature 0.9 for Occitan, greedy decoding for French, and temperature 0.3 for Spanish. Model size alone is not predictive: Aya-8B and even Gemma2-2B outperform larger 14B models, which the paper attributes to better Romance-language representation in pre-training. The study also contributes two newly annotated Medieval Occitan datasets, NAF and Harley, totalling 135,667 tokens.

Load-bearing premise

The gold POS labels for all seven datasets are accurate enough to serve as ground truth, yet no inter-annotator agreement or annotation-quality metric is reported; if those labels contain systematic errors, every accuracy number and comparative conclusion shifts.

Editorial extensions

If this is right

  • When any annotated medieval Romance data exists, researchers should fine-tune rather than prompt; the accuracy gap on NAF is +9.72 points and on Lanfranco +8.63 points.
  • For extremely low-resource varieties, pooling data from related medieval Romance languages is the recommended default, with the expectation of gains like the +5.69 points seen on Harley.
  • Model choice should prioritize Romance-language pre-training over raw size; a 2B model can outperform 14B models on these texts.
  • If fine-tuning is infeasible, few-shot prompting with language-matched decoding (temperature 0.9 for Occitan, greedy for French, temperature 0.3 for Spanish) is the best fallback.
  • Negative transfer is possible, as in Lapidaire's -3.79-point drop, so transfer should be validated per domain before adoption.

Reading between the lines

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

  • The Harley result suggests a testable rule: transfer benefit grows as target-corpus size shrinks, which could be checked by subsampling the larger datasets and plotting gain against training size.
  • The negative transfer on Lapidaire hints that domain vocabulary distance can outweigh language-family closeness; a similar experiment with Anglo-Norman or other non-medical genres would clarify when pooling hurts.
  • Because the paper reports accuracy without annotation-quality metrics, an independent re-annotation of a sample of each corpus would show whether the reported rankings survive gold-label noise.
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

5 major / 5 minor

Summary. This manuscript presents a systematic evaluation of POS tagging for seven low-resource medieval Romance datasets (Medieval Occitan, French, and Spanish) using seven open-weight instruction-tuned LLMs. It compares zero-shot versus few-shot prompting and several decoding strategies, then compares single-dataset fine-tuning against multilingual fine-tuning, which it calls cross-lingual transfer. The paper reports that fine-tuning generally outperforms prompting, that multilingual training yields an average improvement of +0.86 percentage points with the largest gain (+5.69) on the low-resource Harley corpus, that model family and pretraining language coverage matter more than parameter count, and it releases two new Medieval Occitan datasets and an open codebase with detailed results.

Significance. If the empirical claims hold, the paper would make a useful contribution to historical NLP: it introduces two new Medieval Occitan datasets totaling 135,667 tokens, runs a broad experiment matrix across languages, models, and decoding strategies, and releases code and detailed results for reproducibility. The systematic comparison of fine-tuning and prompting on genuinely non-standardized medieval texts is valuable, and the error analysis provides concrete practical guidance. However, the central conclusions are currently undermined by a design confound in the transfer experiments, the absence of any uncertainty estimates, and unmeasured annotation quality in the gold data; these issues are load-bearing for the headline claims, though they appear fixable within the scope of a revision.

major comments (5)
  1. [Section 3.1, Table 1 (Setting b), Table 5] The cross-lingual transfer condition is not actually a transfer condition as implemented. Table 1 defines Setting (b) as fine-tuning on 80% of all datasets combined and evaluating on the held-out 20% of a target dataset; because the target's own 80% training split is part of that combined training set, the +0.86 pp average and the +5.69 pp Harley gain in Table 5 conflate receiving additional in-domain training data (augmentation) with transfer from other languages. To support RQ3, the target dataset must be held out entirely, or the comparison must be between single-dataset training and training on all other datasets while excluding the target's training split; otherwise the headline 'cross-lingual transfer' claim is not supported by this design.
  2. [Tables 3-5, Section 4.2.2] All accuracy numbers come from a single 80/20 split with no repeated seeds or uncertainty estimates, so the paper has no way to distinguish systematic effects from split noise. For example, qwen_14b NAF changes by +0.09 pp, gemma_9b Cauliaco by -0.60 pp, and mistral_7b Lapidaire by -11.45 pp; without variance or significance tests, even large deltas such as Harley +5.69 could be split-specific. The authors should rerun with multiple seeds or splits and report means and standard deviations (or confidence intervals), and temper claims about small differences accordingly.
  3. [Appendix B, Table 7; Section 4.1.1] The condition labeled 'Few-shot' in Table 7 is not few-shot in the standard sense: the prompt contains no labeled POS-tagged examples, only an instruction and two etymological cognate sets (tercia/tersa/tierce/tercera, sanguina/sanc/sang/sangre). Consequently, the conclusion in Section 4.1.1 that 'providing examples during prompting helps the model understand the POS tagging task' is not supported; this comparison shows the effect of an added descriptive instruction, not of in-context examples. Either rename the condition and re-interpret the results, or include actual tagged examples in the few-shot prompt.
  4. [Section 4.2.1, Tables 2 and 3; Section 5] The claim that fine-tuning 'consistently provides more robust performance across language varieties' is contradicted by the Chauliac dataset, where fine-tuning gives 0.8413 against the best few-shot prompting result of 0.8815 (Table 2 vs Table 3). The text acknowledges a 'slight decrease of 0.0402' but the abstract, the recommendations in Section 5, and the conclusion state that fine-tuning consistently outperforms prompting. This qualification must be carried through the paper's central claims, or the claim needs to be restricted to the datasets where it actually holds.
  5. [Section 3.1] The gold annotations are treated as ground truth without any quality measurement. Section 3.1 reports that transcriptions come from HTR models and that POS annotation relies on a modern Occitan tagger with manual corrections, existing editions, or prior annotations, but no inter-annotator agreement, annotation guidelines, or error-rate estimates are provided for any of the seven datasets. If the gold labels contain systematic errors, every reported accuracy and all comparative conclusions are affected; the authors should report at least a sample-based quality assessment or an estimate of annotation noise.
minor comments (5)
  1. [Appendix C, Eq. (1)] The definition of accuracy via TP/TN/FP/FN is unusual for per-token multiclass POS tagging and could confuse readers; please clarify that accuracy is simply the proportion of correctly tagged tokens.
  2. [Table 5 caption] The caption says 'percentual points'; this should be 'percentage points'.
  3. [Table 1] The hardware description says RTX 4090 for '7B-12B' models, but the model suite includes 14B and 2B models; please align the hardware description with the actual model sizes used.
  4. [Tables 3-5; Appendix] The paper does not report fine-tuning hyperparameters (learning rate, number of epochs, LoRA rank or full fine-tuning, sequence length, batch size), which are needed to reproduce Tables 3-5; please provide these details in the appendix.
  5. [Section 5] The recommendation 'always prefer fine-tuning over prompting' repeats the unqualified claim contradicted by the Chauliac results; add the caveat identified in the major comments.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: the accuracy comparisons are direct measurements, and the paper's self-citations are context or data provenance rather than load-bearing evidence.

full rationale

The paper is an empirical benchmark study; it derives no formal predictions from fitted parameters and contains no first-principles result that is equivalent to its inputs by construction. The headline comparisons—fine-tuning versus prompting (Section 4.2.1, Tables 2–3) and multilingual versus single-dataset fine-tuning (Tables 4–5)—are direct accuracy measurements on held-out 20% splits, and the reported deltas are arithmetic differences of those measurements, not fitted parameters renamed as predictions. The only self-citations appear as related-work context (Schöffel et al., 2025) and as dataset transcription provenance (Wiedner, 2023); neither carries the paper's central argument, so there is no self-citation chain forcing a conclusion. The cross-lingual transfer condition (Setting (b), Table 1) trains on 80% of all datasets combined, including the target dataset's own training split, so the +0.86 percentage-point average and +5.69 Harley gain conflate additional in-domain training data with transfer to an unseen variety; this is a genuine experimental-design and interpretation limitation, and the absence of variance estimates further weakens the strength of the claims, but it does not make the reported numbers equivalent to their inputs by definition, and no fitted parameter is being presented as an independent prediction. Section 5's recommendations are explicitly summaries of the measured patterns rather than out-of-sample predictions. No circular step meeting the quoted-evidence threshold was found.

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

This is an empirical benchmarking paper, so the claims rest on data quality and evaluation protocol rather than fitted model parameters. No new physical or formal entities are postulated. The key unstated premises are the accuracy of the gold labels, the adequacy of single-run evaluations, and the appropriateness of accuracy as the headline metric.

assumptions (3)
  • domain assumption Gold POS annotations for all seven datasets are accurate enough to serve as ground truth.
    Section 3.1 describes transcription and manual correction, but no inter-annotator agreement or quality measures are reported; all accuracy numbers inherit this assumption.
  • domain assumption Accuracy is a sufficient metric for the study's conclusions.
    Section 4 and Appendix D focus on accuracy; the paper notes in Limitations that tag-specific variations may be obscured, and Figure 2 shows imbalanced tag distributions.
  • domain assumption Single-run evaluation results are stable enough to support the reported comparisons.
    No repeated runs, confidence intervals, or significance tests are reported (Tables 2-5); differences as small as +0.0014 are treated as meaningful.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Unveiling Factors for Enhanced POS Tagging: A Study of Low-Resource Medieval Romance Languages." pith.science (2026). https://pith.science/paper/RVAIF7DZ

@misc{pith2026250617715,
  author       = {Pith},
  title        = {Pith review of: Unveiling Factors for Enhanced POS Tagging: A Study of Low-Resource Medieval Romance Languages},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RVAIF7DZ}},
  note         = {Machine review of arXiv:2506.17715}
}
read the original abstract

Part-of-speech (POS) tagging remains a foundational component in natural language processing pipelines, particularly critical for historical text analysis at the intersection of computational linguistics and digital humanities. Despite significant advancements in modern large language models (LLMs) for ancient languages, their application to Medieval Romance languages presents distinctive challenges stemming from diachronic linguistic evolution, spelling variations, and labeled data scarcity. This study systematically investigates the central determinants of POS tagging performance across diverse corpora of Medieval Occitan, Medieval Spanish, and Medieval French texts, spanning biblical, hagiographical, medical, and dietary domains. Through rigorous experimentation, we evaluate how fine-tuning approaches, prompt engineering, model architectures, decoding strategies, and cross-lingual transfer learning techniques affect tagging accuracy. Our results reveal both notable limitations in LLMs' ability to process historical language variations and non-standardized spelling, as well as promising specialized techniques that effectively address the unique challenges presented by low-resource historical languages.

Figures

Figures reproduced from arXiv: 2506.17715 by the authors.

Figure 1
Figure 1. Romance languages in medieval texts: geographical distribution of (a) French and Occitan [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Normalized Part-of-Speech distribution for the corpus: Albucasis (blue), Vida de Sant [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 32 canonical work pages

  1. [1]

    Hewett, Mojan Javaheripi, Piero Kauffmann, James R

    Marah Abdin, Jyoti Aneja, Harkirat Behl, Sébastien Bubeck, Ronen Eldan, Suriya Gunasekar, Michael Harrison, Russell J. Hewett, Mojan Javaheripi, Piero Kauffmann, James R. Lee, Yin Tat Lee, Yuanzhi Li, Weishung Liu, Caio C. T. Mendes, Anh Nguyen, Eric Price, Gustavo de Rosa, Olli Saarikivi, Adil Salim, Shital Shah, Xin Wang, Rachel Ward, Yue Wu, Dingli Yu,...

    arXiv 2024

  2. [2]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023

    arXiv 2023

  3. [3]

    A learning algorithm for boltzmann machines

    David H Ackley, Geoffrey E Hinton, and Terrence J Sejnowski. A learning algorithm for boltzmann machines. Cognitive science, 9 0 (1): 0 147--169, 1985

    1985

  4. [4]

    Improving Zero-shot Cross-lingual Transfer between Closely Related Languages by injecting Character-level Noise

    Noëmi Aepli and Rico Sennrich. Improving zero-shot cross-lingual transfer between closely related languages by injecting character-level noise, 2022. URL https://arxiv.org/abs/2109.06772

    work page Pith review arXiv 2022

  5. [5]

    Part-of-Speech Tagging on an Endangered Language: a Parallel Griko-Italian Resource

    Antonis Anastasopoulos, Marika Lekakou, Josep Quer, Eleni Zimianiti, Justin DeBenedetto, and David Chiang. Part-of-speech tagging on an endangered language: a parallel griko-italian resource, 2018. URL https://arxiv.org/abs/1806.03757

    work page Pith review arXiv 2018

  6. [6]

    A Falta de Pan, Buenas Son Tortas: The Efficacy of Predicted UPOS Tags for Low Resource UD Parsing

    Mark Anderson, Mathieu Dehouck, and Carlos Gómez Rodríguez. A falta de pan, buenas son tortas: The efficacy of predicted upos tags for low resource ud parsing, 2021. URL https://arxiv.org/abs/2106.04222

    work page Pith review arXiv 2021

  7. [7]

    Aya 23: Open weight releases to further multilingual progress, 2024

    Viraat Aryabumi, John Dang, Dwarak Talupuru, Saurabh Dash, David Cairuz, Hangyu Lin, Bharat Venkitesh, Madeline Smith, Jon Ander Campos, Yi Chern Tan, Kelly Marchisio, Max Bartolo, Sebastian Ruder, Acyr Locatelli, Julia Kreutzer, Nick Frosst, Aidan Gomez, Phil Blunsom, Marzieh Fadaee, Ahmet Üstün, and Sara Hooker. Aya 23: Open weight releases to further m...

    arXiv 2024

  8. [8]

    How Low is Too Low? A Computational Perspective on Extremely Low-Resource Languages

    Rachit Bansal, Himanshu Choudhary, Ravneet Punia, Niko Schenk, Jacob L Dahl, and Émilie Pagé-Perron. How low is too low? a computational perspective on extremely low-resource languages, 2021. URL https://arxiv.org/abs/2105.14515

    work page Pith review arXiv 2021

Show all 52 references
  1. [9]

    Does manipulating tokenization aid cross-lingual transfer? a study on pos tagging for non-standardized languages, 2023

    Verena Blaschke, Hinrich Schütze, and Barbara Plank. Does manipulating tokenization aid cross-lingual transfer? a study on pos tagging for non-standardized languages, 2023. URL https://arxiv.org/abs/2304.10158

    2023 arXiv

  2. [10]

    A large-scale comparison of historical text normalization systems

    Marcel Bollmann. A large-scale comparison of historical text normalization systems. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, 1: 0 3885--3898, 2019

    2019

  3. [11]

    Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al

    Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in Neural Information Processing Systems, 33: 0 1877--1901, 2020

    1901

  4. [12]

    Corpus and models for lemmatisation and pos-tagging of old french, 2021

    Jean-Baptiste Camps, Thibault Clérice, Frédéric Duval, Lucence Ing, Naomi Kanaoka, and Ariane Pinche. Corpus and models for lemmatisation and pos-tagging of old french, 2021. URL https://arxiv.org/abs/2109.11442

    2021 arXiv

  5. [13]

    Low-resource name tagging learned with weakly labeled data, 2019

    Yixin Cao, Zikun Hu, Tat-Seng Chua, Zhiyuan Liu, and Heng Ji. Low-resource name tagging learned with weakly labeled data, 2019. URL https://arxiv.org/abs/1908.09659

    2019 arXiv

  6. [14]

    A grounded unsupervised universal part-of-speech tagger for low-resource languages, 2019

    Ronald Cardenas, Ying Lin, Heng Ji, and Jonathan May. A grounded unsupervised universal part-of-speech tagger for low-resource languages, 2019. URL https://arxiv.org/abs/1904.05426

    2019 arXiv

  7. [15]

    External history of french

    Philippe Caron. External history of french. In Wendy Ayres-Bennett and Mairi McLaughlin (eds.), The Oxford handbook of the French language, chapter 4, pp.\ 143–162. Oxford University Press, Oxford, 2024

    2024

  8. [16]

    Reducing confusion in active learning for part-of-speech tagging, 2020

    Aditi Chaudhary, Antonios Anastasopoulos, Zaid Sheikh, and Graham Neubig. Reducing confusion in active learning for part-of-speech tagging, 2020. URL https://arxiv.org/abs/2011.00767

    2020 arXiv

  9. [17]

    Zero resource cross-lingual part of speech tagging, 2024

    Sahil Chopra. Zero resource cross-lingual part of speech tagging, 2024. URL https://arxiv.org/abs/2401.05727

    2024 arXiv

  10. [18]

    The importance of context in very low resource language modeling, 2022

    Lukas Edman, Antonio Toral, and Gertjan van Noord. The importance of context in very low resource language modeling, 2022. URL https://arxiv.org/abs/2205.04810

    2022 arXiv

  11. [19]

    Beam search strategies for neural machine translation

    Markus Freitag and Yaser Al-Onaizan. Beam search strategies for neural machine translation. In Proceedings of the First Workshop on Neural Machine Translation. Association for Computational Linguistics, 2017. doi:10.18653/v1/w17-3207. URL http://dx.doi.org/10.18653/v1/W17-3207

    2017 doi

  12. [20]

    From freem to d'alembert: a large corpus and a language model for early modern french, 2022

    Simon Gabay, Pedro Ortiz Suarez, Alexandre Bartz, Alix Chagué, Rachel Bawden, Philippe Gambette, and Benoît Sagot. From freem to d'alembert: a large corpus and a language model for early modern french, 2022. URL https://arxiv.org/abs/2202.09452

    2022 arXiv

  13. [21]

    Cirugía mayor

    Francisco Gago Jover. Cirugía mayor. In Spanish Medical Texts. Digital Library of Old Spanish Texts. Hispanic Seminary of Medieval Studies, 2011 a . URL http://www.hispanicseminary.org/t&c/ac/index.htm

    2011

  14. [22]

    Tratado de cirugía

    Francisco Gago Jover. Tratado de cirugía. In Spanish Medical Texts. Digital Library of Old Spanish Texts. Hispanic Seminary of Medieval Studies, 2011 b . URL http://www.hispanicseminary.org/t&c/ac/index.htm

    2011

  15. [23]

    Gemma-Team, Morgane Riviere, Shreya Pathak, Pier Giuseppe Sessa, Cassidy Hardin, Surya Bhupatiraju, Léonard Hussenot, Thomas Mesnard, Bobak Shahriari, Alexandre Ramé, Johan Ferret, Peter Liu, Pouya Tafti, Abe Friesen, Michelle Casbon, Sabela Ramos, Ravin Kumar, Charline Le Lan...

    2024 arXiv

  16. [25]

    The llama 3 herd of models

    Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, et al. The llama 3 herd of models. arXiv preprint arXiv:2407.21783, 2024 b

    2024 arXiv

  17. [26]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning

    Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948, 2025

    2025 arXiv

  18. [27]

    The curious case of neural text degeneration

    Ari Holtzman, Jan Buys, Li Du, Maxwell Forbes, and Yejin Choi. The curious case of neural text degeneration. arXiv preprint arXiv:1904.09751, 2019

    1904 arXiv

  19. [28]

    Graph-based multilingual label propagation for low-resource part-of-speech tagging, 2022

    Ayyoob Imani, Silvia Severini, Masoud Jalili Sabet, François Yvon, and Hinrich Schütze. Graph-based multilingual label propagation for low-resource part-of-speech tagging, 2022. URL https://arxiv.org/abs/2210.09840

    2022 arXiv

  20. [29]

    Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Thomas ...

    2023

  21. [30]

    Weakly supervised pos taggers perform poorly on truly low-resource languages, 2020

    Katharina Kann, Ophélie Lacroix, and Anders Søgaard. Weakly supervised pos taggers perform poorly on truly low-resource languages, 2020. URL https://arxiv.org/abs/2004.13305

    2020 arXiv

  22. [31]

    Topro: Token-level prompt decomposition for cross-lingual sequence labeling tasks, 2024

    Bolei Ma, Ercong Nie, Shuzhou Yuan, Helmut Schmid, Michael Färber, Frauke Kreuter, and Hinrich Schütze. Topro: Token-level prompt decomposition for cross-lingual sequence labeling tasks, 2024. URL https://arxiv.org/abs/2401.16589

    2024 arXiv

  23. [32]

    Krzysztof Nowak, Jędrzej Ziębura, Krzysztof Wróbel, and Aleksander Smywiński-Pohl. efontes. part of speech tagging and lemmatization of medieval latin texts.a cross-genre survey, 2024. URL https://arxiv.org/abs/2407.00418

    2024 arXiv

  24. [33]

    Natural Language Processing for Historical Texts, volume 5 of Synthesis Lectures on Human Language Technologies

    Michael Piotrowski. Natural Language Processing for Historical Texts, volume 5 of Synthesis Lectures on Human Language Technologies. Morgan & Claypool Publishers, 2012

    2012

  25. [34]

    The best of both worlds: Lexical resources to improve low-resource part-of-speech tagging, 2018

    Barbara Plank, Sigrid Klerke, and Zeljko Agic. The best of both worlds: Lexical resources to improve low-resource part-of-speech tagging, 2018. URL https://arxiv.org/abs/1811.08757

    2018 arXiv

  26. [35]

    Distant supervision from disparate sources for low-resource part-of-speech tagging, 2018

    Barbara Plank and Željko Agić. Distant supervision from disparate sources for low-resource part-of-speech tagging, 2018. URL https://arxiv.org/abs/1808.09733

    2018 arXiv

  27. [36]

    Ponti, Ivan Vulić, Ryan Cotterell, Marinela Parovic, Roi Reichart, and Anna Korhonen

    Edoardo M. Ponti, Ivan Vulić, Ryan Cotterell, Marinela Parovic, Roi Reichart, and Anna Korhonen. Parameter space factorization for zero-shot learning across tasks and languages, 2020. URL https://arxiv.org/abs/2001.11453

    2020 arXiv

  28. [37]

    La linguistique outillée à l'épreuve de la variation : Ressources pour l'analyse de parlers occitans de l'Ariège

    Clamenca Poujade. La linguistique outillée à l'épreuve de la variation : Ressources pour l'analyse de parlers occitans de l'Ariège. PhD thesis, Université de Toulouse, In progress

  29. [38]

    Profiterole : un corpus morpho-syntaxique et syntaxique de fran c ais m \'e di \'e val

    Sophie Pr \'e vost, Lo \"i c Grobol, Mathieu Dehouck, Alexei Lavrentiev, and Serge Heiden. Profiterole : un corpus morpho-syntaxique et syntaxique de fran c ais m \'e di \'e val . Corpus , 0 (25): 0 8538, January 2024. doi:10.4000/corpus.8538. URL https://hal.science/hal-04681591

    2024 doi

  30. [39]

    Qwen2.5 technical report, 2025

    Qwen-Team, :, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, Junyang Lin, Kai Dang, Keming Lu, Keqin Bao, Kexin Y...

    2025 arXiv

  31. [40]

    Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research, 21 0 (140): 0 1--67, 2020

    2020

  32. [41]

    Heidelberg-boston @ sigtyp 2024 shared task: Enhancing low-resource language analysis with character-aware hierarchical transformers, 2024

    Frederick Riemenschneider and Kevin Krahn. Heidelberg-boston @ sigtyp 2024 shared task: Enhancing low-resource language analysis with character-aware hierarchical transformers, 2024. URL https://arxiv.org/abs/2405.20145

    2024 arXiv

  33. [42]

    Modern models, medieval texts: A pos tagging study of old occitan, 2025

    Matthias Schöffel, Marinus Wiedner, Esteban Garces Arias, Paula Ruppert, Christian Heumann, and Matthias Aßenmacher. Modern models, medieval texts: A pos tagging study of old occitan, 2025. URL https://arxiv.org/abs/2503.07827

    2025 arXiv

  34. [43]

    Lapidaire en prose

    Paul Studer and Joan Evans (eds.). Lapidaire en prose. Champion, Paris, 1624

  35. [44]

    High-resource methodological bias in low-resource investigations, 2022

    Maartje ter Hoeve, David Grangier, and Natalie Schluter. High-resource methodological bias in low-resource investigations, 2022. URL https://arxiv.org/abs/2211.07534

    2022 arXiv

  36. [45]

    anathomie

    Sabine Tittel. Die "anathomie" in der "grande chirurgie" des gui de chauliac : wort- und sachgeschichtliche untersuchungen und edition, 2004

    2004

  37. [46]

    Recipe for zero-shot pos tagging: Is it useful in realistic scenarios?, 2024

    Zeno Vandenbulcke, Lukas Vermeire, and Miryam de Lhoneux. Recipe for zero-shot pos tagging: Is it useful in realistic scenarios?, 2024. URL https://arxiv.org/abs/2410.10576

    2024 arXiv

  38. [47]

    Old O ccitan handwriting

    Marinus Wiedner. Old O ccitan handwriting. (modell-nr. 52822, CER =3,51\ P y L aia- M odell for handwritten O ccitan from the 13th and 14th century., 2023. URL https://readcoop.eu/model/old-occitan-handwriting/

    2023

  39. [48]

    Natural language processing for similar languages, varieties, and dialects: A survey

    Marcos Zampieri, Shervin Malmasi, Yves Scherrer, Tanja Samard z i\' c , Francis Tyers, Miikka Silfverberg, Natalia Klyueva, Tung-Le Pan, Chu-Ren Huang, Radu Tudor Ionescu, et al. Natural language processing for similar languages, varieties, and dialects: A survey. Natural Lang...

    2019

  40. [49]

    Bridging pre-trained language models and hand-crafted features for unsupervised pos tagging, 2022

    Houquan Zhou, Yang Li, Zhenghua Li, and Min Zhang. Bridging pre-trained language models and hand-crafted features for unsupervised pos tagging, 2022. URL https://arxiv.org/abs/2203.10315

    2022 arXiv

  41. [50]

    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...

  42. [51]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  43. [52]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  44. [53]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.