Pith. sign in

REVIEW 4 major objections 5 minor 65 references

Machine-translated Russian corpora can supplement scarce Finnish CEFR training data, this paper argues.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review

2026-08-01 13:25 UTC pith:NWBHZJK4

load-bearing objection The paper's own Table 5 contradicts its headline claim: native-only training beats native+MT on the native Finnish test set. the 4 major comments →

arxiv 2607.19101 v1 pith:NWBHZJK4 submitted 2026-07-21 cs.CL cs.LG

Translation as Augmentation: Effect of Translated Data on Assessment of Difficulty

Yiheng Wu , Jue Hou , Roman Yangarber This is my paper
classification cs.CL cs.LG
keywords text difficulty assessmentCEFRmachine translation augmentationlow-resource NLPFinnishRussianBERT regressionreadability
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 tackles the shortage of expert-labeled CEFR data in Finnish by translating Russian texts that already carry difficulty labels and adding them to the training set. It trains BERT-based regressors to predict a continuous difficulty score, and reports that combining native Finnish with translated Russian data improves robustness and accuracy compared with training on either alone. On the paper's own results, the clearest gain is on machine-translated test data; on native Finnish test text the combined model is close to, but slightly below, a native-only model. The intended contribution is a recipe for low-resource languages that lack expert annotations.

Core claim

The paper claims that cross-lingual data augmentation via machine translation transfers CEFR difficulty labels from Russian to Finnish well enough to train a useful difficulty estimator. Using 3,460 native Finnish documents and 7,457 translated documents, a fine-tuned BERT regression model reaches R²=96.01% on native Finnish when trained on both, versus 97.30% for native-only training, and reaches R²=94.26% on machine-translated Finnish text, where native-only training collapses. The author's reading is that MT data supplements scarce native data and regularizes the model; the paper also reports that mid-level (B1-B2) translated texts help most, and that 40% proportional MT sampling is as ef

What carries the argument

An augmentation pipeline that maps 11 CEFR sub-levels to a continuous 1-6 scale and uses an off-the-shelf Russian-to-Finnish MT system to create synthetic Finnish texts carrying Russian difficulty labels. The translated corpus triples the training data; a BERT regression head then learns to predict the score. The assumption that the MT system preserves difficulty level across translation is the mechanism that makes the labels usable.

Load-bearing premise

The entire method rests on the assumption that the machine translation system preserves the CEFR difficulty level of a Russian text when rendering it in Finnish; the paper validates this only by manual inspection of a sample, not by any quantitative check.

What would settle it

If a quantitative comparison of source Russian CEFR labels with labels assigned to the Finnish translations by human experts (or by a strong Finnish difficulty model) showed a systematic shift in level for more than a small fraction of texts, the augmented labels would be unreliable and the reported gains would be artifacts. A direct check: take a held-out set of Russian texts with known CEFR levels, translate them, have expert raters label the Finnish versions, and measure the confusion matrix.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • If true, low-resource languages can bootstrap difficulty models by translating existing high-resource annotated corpora rather than commissioning new expert annotations.
  • Translated data alone is insufficient for native-language prediction; some native data is still required, so MT augmentation is complementary, not a replacement.
  • Mid-level CEFR texts transfer best, so selective sampling by source difficulty should be part of any augmentation strategy.
  • A difficulty model trained this way could act as a critic in LLM simplification pipelines and personalized learning tools for Finnish.
  • Proportional augmentation around 40% translated data gives accuracy comparable to full augmentation, suggesting a saturation point beyond which extra MT data adds noise.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper's claim that augmentation 'significantly improves' accuracy is not supported by the native-test comparison in Table 5; the improvement appears only when the test set is itself machine-translated, so the practical benefit may depend on how much translated text the deployed system will see.
  • A stronger validation would test on Finnish texts translated by humans from Russian, or on native Finnish with known original Russian labels, to separate language-transfer effects from label-preservation artifacts.
  • The B1-B2 dominance suggests a label-preservation filter: training examples whose source-level or round-trip translation changes difficulty should be down-weighted or excluded.
  • One could use the same pipeline in reverse (Finnish to Russian) or on another language pair to check whether the result is specific to Russian-Finnish or general.

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

4 major / 5 minor

Summary. The paper proposes using machine-translated (MT) Russian texts as synthetic augmented training data for Finnish text-difficulty assessment, where native annotated Finnish data are scarce. It trains BERT-based regression models over native Finnish, MT-translated Russian, and combined training sets, reporting MSE, MAE, and R² on native and MT test sets. The central claim, as stated in the abstract and Section 4, is that augmenting scarce native data with MT data significantly improves difficulty-estimation accuracy for Finnish. The paper also explores ablations with different proportions and CEFR-level subsets of MT data.

Significance. If the central claim were true, the approach would offer a practical way to bootstrap difficulty-assessment systems for low-resource languages by leveraging existing Russian annotations. The experimental matrix is fairly broad, including multiple train/test configurations and ablations, and the paper is candid about some of its assumptions (§3.2) and limitations (§6). However, the paper's own results contradict the headline: the best performance on native Finnish test data comes from native-only training, not from MT augmentation. The only large gains from augmentation occur when the test set is itself machine-translated, which is a distribution-matching artifact rather than evidence of generalization. The lack of quantitative validation of the central label-preservation assumption further weakens the empirical chain. The paper does not provide code, data, or machine-checked artifacts, so reproducibility is limited.

major comments (4)
  1. [Abstract; §4.2; Table 5] The abstract claims that augmenting scarce native data with machine-translated corpora significantly improves accuracy. Table 5 directly contradicts this on the native Finnish test set: Exp. 1 (FI native-only training) gives MSE=5.12 and R²=97.30, while Exp. 5 (FI native+MT training) gives MSE=7.57 and R²=96.01. On the evaluation that matters for the stated application (native input texts), adding MT data degrades performance. Section 4.2 acknowledges Exp. 1 as 'strong in-domain performance' but then describes the combined model as 'comparable accuracy' and concludes that translation-based augmentation is effective. The reported numbers do not support the abstract's unqualified claim.
  2. [§4.2; Table 5, Exp. 6] The result highlighted as a 'dramatic performance gain' (MSE=4.22, R²=94.26) is obtained by training on native+MT and testing on MT-translated texts generated by the same OpusMT pipeline used to create the augmented training data. This is a distribution-overlap effect: the test set shares the same translation artifacts and label-injection procedure as the training set. It provides no evidence for generalization to real-world native Finnish input, and therefore cannot support the paper's conclusion that MT augmentation improves low-resource Finnish difficulty assessment.
  3. [§3.2] The paper explicitly states that machine translation does not guarantee preservation of CEFR difficulty level across languages, and relies only on 'manual inspection of a sample by native language experts' to assert that the OpusMT model preserves difficulty. No sample size, inter-annotator agreement, or quantitative validation is reported. If CEFR levels are not preserved, every augmented label is potentially wrong, invalidating the training data in Experiments 1–6 and 9–16. This is a load-bearing assumption for both research questions and is not adequately supported.
  4. [§4.3; Table 6] The ablation section interprets Exp. 10 (B1–B2 injection, R²=96.68) and Exp. 13 (40% MT, R²=96.20) as 'strongest improvement' and 'slightly outperforming' full augmentation. However, both are still below the native-only baseline of Exp. 1 (R²=97.30) on the same native test set. The omitted comparison against the native-only baseline makes the narrative in §4.3 misleading and further weakens the claim that MT data helps.
minor comments (5)
  1. [§3.1; Tables 1 and 3] The text states the native corpus contains 4544 texts, while Table 1 and Table 3 report a native total of 3460. Similarly, the text says the total dataset contains 12,865 texts, but Table 1 reports 10,917. Please reconcile these numbers.
  2. [Table 5 caption] Typo: 'mean quared error' should be 'mean squared error'.
  3. [§4.2; Figure 1] References such as 'Figure 1.4' and 'Figure 1.5' are unusual; standard notation would be 'Figure 1(d)' or 'Figure 1, Exp. 4'.
  4. [§4.2] The phrase 'while applying dataset balancing' is vague. The balancing procedure, its parameters, and its effect on the comparison should be described explicitly, since it is a possibly important confound.
  5. [Table 4] The table lists six linguistic features but the text notes they are not used in modeling. Consider clarifying their role (descriptive only) to avoid confusion.

Circularity Check

1 steps flagged

Central augmentation claim is supported by an MT-test evaluation whose labels inherit the same source-label projection used to construct the training data; the independent native test contradicts the claim.

specific steps
  1. self definitional [Section 3.2 (Translated Data); Section 4.2, Table 5, Exp 6]
    "machine translation does not guarantee that a text in Russian will remain at the same difficulty level after translation into Finnish. ... However, combining translated and native Finnish data—while applying dataset balancing—yields a dramatic performance gain, reaching MSE of 4.22%, MAE of 8.24%, and R2 of 94.26%."

    The MT test set's labels are the same source-inherited CEFR labels used to construct the MT training data, and the paper explicitly leaves difficulty preservation across MT unvalidated ('does not guarantee... confirmed by manual inspection of a sample'). Exp6 therefore measures how well the model reproduces the label-construction rule on in-pipeline MT text, not native Finnish difficulty. The abstract's unqualified 'significantly improves' relies on this self-consistent evaluation, while the independent native test shows augmentation degrades performance (Exp1: MSE 5.12, R2 97.30 vs Exp5: MSE 7.57, R2 96.01).

full rationale

No derivation chain or equations are present, so most circularity patterns (fitted-input-as-prediction, self-citation uniqueness, ansatz-by-citation, renaming) do not apply. The paper's own data sources include prior work by the same group (Katinskaia et al. 2025; Dmitrieva 2025), but citing those resources is ordinary scientific practice, not load-bearing circularity. The main circular component is the evaluation on machine-translated test data: labels for MT-translated Finnish texts are defined as the Russian source labels, and the same label-projection pipeline generated the augmented training data. Success on that test is therefore partially self-consistent with the training-label construction. However, the paper also reports a native Finnish test that is independent of this construction, and that test contradicts the claimed benefit. The abstract's overstatement is a correctness and reporting problem rather than a fully circular derivation; the circular MT-test evidence inflates the appearance of success, but the independent baseline prevents a score above 4. The section 3.2 admission about unverified label preservation is accurately flagged as a limitation, but it is not itself a circular step.

Axiom & Free-Parameter Ledger

3 free parameters · 3 axioms · 0 invented entities

The paper introduces no new entities. Its central claim rests on hand-made labeling decisions: linearizing CEFR and transferring Russian annotations through MT without independent validation. The unspecified balancing and outlier-removal procedures further make the dataset construction non-exhaustive.

free parameters (3)
  • CEFR-to-score linear mapping = A1=1.0, A1+=1.5, A2=2.0, A2+=2.5, B1=3.0, B1+=3.5, B2=4.0, B2+=4.5, C1=5.0, C1+=5.5, C2=6.0
    Section 3.1: 'we make the assumption that the CEFR levels are evenly spaced on a linear difficulty scale' because exact spacing is latent; the regression target depends on this mapping.
  • Russian corpus outlier removal = unspecified
    Section 3.2: 'we removed some outlier instances' with no thresholds; this changes the MT counts and split sizes without a documented rule.
  • Dataset balancing procedure = unspecified
    Section 4.2 says the combined model uses 'dataset balancing' but does not describe the procedure, making the reported gain non-reproducible.
axioms (3)
  • ad hoc to paper CEFR levels are evenly spaced on a linear difficulty scale
    Section 3.1 makes this assumption with no external evidence; the regression target 1-6 is entirely defined by this choice.
  • domain assumption Machine translation preserves the CEFR difficulty level of Russian source texts in Finnish
    Section 3.2 concedes MT does not guarantee preservation and offers only manual inspection of a sample; no quantitative evaluation.
  • ad hoc to paper Newspaper texts can be assigned a coarse level 'C' between C1 and C2
    Section 3.1: HS and YLE texts are labeled C1+ based on the assumption that news is approximately between C1 and C2; this affects 1,418 native labels.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Translation as Augmentation: Effect of Translated Data on Assessment of Difficulty." pith.science (2026). https://pith.science/paper/NWBHZJK4

@misc{pith2026260719101,
  author       = {Pith},
  title        = {Pith review of: Translation as Augmentation: Effect of Translated Data on Assessment of Difficulty},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NWBHZJK4}},
  note         = {Machine review of arXiv:2607.19101}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Reliable Text Difficulty Assessment is a prerequisite for valid text simplification workflows and personalized learning applications. However, the development of robust assessment models is severely hindered by a critical bottleneck: the scarcity of expert-annotated corpora containing fine-grained difficulty levels (e.g., CEFR), particularly for lower-resource languages. This paper addresses this data scarcity problem in the context of a low-resource European language. We propose a cross-lingual data augmentation strategy that leverages machine translation to transfer labeled resources from high-resource languages to the target low-resource language. We train BERT-based regression models to predict difficulty scores and investigate whether synthetic, translated data can effectively supplement native training sets. Our experiments demonstrate that augmenting scarce native data with machine-translated corpora significantly improves the accuracy of difficulty estimation, offering a viable solution for languages lacking extensive expert annotations.

Figures

Figures reproduced from arXiv: 2607.19101 by Jue Hou, Roman Yangarber, Yiheng Wu.

Figure 1
Figure 1. Figure 1: Prediction of difficulty, corresponding to experiments in Table [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Box plots for experiments (Exp. 9–16) showing error distributions across models trained with [PITH_FULL_IMAGE:figures/full_fig_p007_2.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

65 extracted references · 8 canonical work pages

  1. [1]

    Aho and Jeffrey D

    Alfred V. Aho and Jeffrey D. Ullman , title =. 1972

    1972

  2. [2]

    Publications Manual , year = "1983", publisher =

    1983

  3. [3]

    Chandra and Dexter C

    Ashok K. Chandra and Dexter C. Kozen and Larry J. Stockmeyer , year = "1981", title =. doi:10.1145/322234.322243

    arXiv 1981

  4. [4]

    Scalable training of

    Andrew, Galen and Gao, Jianfeng , booktitle=. Scalable training of

  5. [5]

    Dan Gusfield , title =. 1997

    1997

  6. [6]

    Tetreault , title =

    Mohammad Sadegh Rasooli and Joel R. Tetreault , title =. Computing Research Repository , volume =. 2015 , url =

    2015

  7. [7]

    A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data , Volume =

    Ando, Rie Kubota and Zhang, Tong , Issn =. A Framework for Learning Predictive Structures from Multiple Tasks and Unlabeled Data , Volume =. Journal of Machine Learning Research , Month = dec, Numpages =

  8. [8]

    Journal of Applied Psychology , year =

    Flesch, Rudolf , title =. Journal of Applied Psychology , year =

  9. [9]

    Estimation of Text Difficulty in the Context of Language Learning

    Katinskaia, Anisia and Vu, Anh-Duc and Hou, Jue and Vanhatalo, Ulla and Wu, Yiheng and Yangarber, Roman. Estimation of Text Difficulty in the Context of Language Learning. BEA: 20th Workshop on Innovative Use of NLP for Building Educational Applications. 2025. doi:10.18653/v1/2025.bea-1.43

    work page doi:10.18653/v1/2025.bea-1.43 2025

  10. [10]

    Findings of the TSAR 2025 Shared Task on Readability-Controlled Text Simplification

    Alva-Manchego, Fernando and Stodden, Regina and Imperial, Joseph Marvin and Barayan, Abdullah and North, Kai and Tayyar Madabushi, Harish. Findings of the TSAR 2025 Shared Task on Readability-Controlled Text Simplification. Proceedings of the Fourth Workshop on Text Simplification, Accessibility and Readability (TSAR 2025). 2025. doi:10.18653/v1/2025.tsar-1.8

    doi:10.18653/v1/2025.tsar-1.8 2025

  11. [11]

    2001 , publisher =

    Common European Framework of Reference for Languages: Learning, teaching, assessment , author =. 2001 , publisher =

    2001

  12. [12]

    Frontiers in Artificial Intelligence , volume=

    Lexical simplification benchmarks for English, Portuguese, and Spanish , author=. Frontiers in Artificial Intelligence , volume=. 2022 , publisher=. doi:10.3389/frai.2022.991242 , url=

    arXiv 2022

  13. [13]

    Tanprasert, Teerapaun and Kauchak, David , booktitle=

  14. [14]

    Derivation of new readability formulas (automated readability index, fog count and

    Kincaid, J Peter and Fishburne Jr, Robert P and Rogers, Richard L and Chissom, Brad S , year=. Derivation of new readability formulas (automated readability index, fog count and

  15. [15]

    Information for all: European standards for making information easy to read and understand , author =

  16. [16]

    Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) , month = nov, year =

    Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation , author =. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) , month = nov, year =

    2020

  17. [17]

    Frontiers in Artificial Intelligence , VOLUME=

    Beauchemin, David and Saggion, Horacio and Khoury, Richard , TITLE=. Frontiers in Artificial Intelligence , VOLUME=. 2023 , URL=. doi:10.3389/frai.2023.1223924 , ISSN=

    arXiv 2023

  18. [18]

    Language Learning , year =

    Nahatame, Shingo and Yamaguchi, Katsuyoshi , title =. Language Learning , year =. doi:10.1111/lang.70011 , url =

    work page doi:10.1111/lang.70011

  19. [19]

    Pattern Recognition Letters , volume=

    Rank consistent ordinal regression for neural networks with application to age estimation , author=. Pattern Recognition Letters , volume=. 2020 , publisher=

    2020

  20. [20]

    Hurst, Aaron and Lerer, Adam and Goucher, Adam P and Perelman, Adam and Ramesh, Aditya and Clark, Aidan and Ostrow, AJ and Welihinda, Akila and Hayes, Alan and Radford, Alec and others , journal=

  21. [21]

    A corpus of

    Laposhina, Antonina , journal=. A corpus of

  22. [22]

    Creating an Aligned R ussian Text Simplification Dataset from Language Learner Data

    Dmitrieva, Anna and Tiedemann, J. Creating an Aligned R ussian Text Simplification Dataset from Language Learner Data. Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing. 2021

    2021

  23. [23]

    Towards Fine-Grained Pedagogical Control over English Grammar Complexity in Educational Text Generation

    Glandorf, Dominik and Meurers, Detmar. Towards Fine-Grained Pedagogical Control over English Grammar Complexity in Educational Text Generation. Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024). 2024

    2024

  24. [24]

    Akram and Aziz, Abdul and Faruqui, Rokan Uddin and Chy, Abu Nowshed

    Ayman, Nabila and Hossain, Md. Akram and Aziz, Abdul and Faruqui, Rokan Uddin and Chy, Abu Nowshed. BengaliLCP: A Dataset for Lexical Complexity Prediction in the Bengali Texts. Proceedings of LREC-COLING: the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation. 2024

    2024

  25. [25]

    Determining Perceived Text Complexity: An Evaluation of German Sentences Through Student Assessments

    Boris Thome and Friederike Hertweck and Stefan Conrad. Determining Perceived Text Complexity: An Evaluation of German Sentences Through Student Assessments. Proceedings of the 17th International Conference on Educational Data Mining, EDM 2024. 2024

    2024

  26. [26]

    Complexity of German Texts Written by Primary School Children

    Laâguidi, Jammila and Neumann, Dana and Laarmann-Quante, Ronja and Dipper, Stefanie and Chifligarov, Mihail. Complexity of German Texts Written by Primary School Children. Proceedings of the 20th Conference on Natural Language Processing (KONVENS 2024). 2024

    2024

  27. [27]

    Chinese Automatic Readability Assessment Using Adaptive Pre-training and Linguistic Feature Fusion

    Yang, Xusheng and Yang, Jincai and Li, Xiao. Chinese Automatic Readability Assessment Using Adaptive Pre-training and Linguistic Feature Fusion. Proceedings of the 31st International Conference on Computational Linguistics. 2025

    2025

  28. [28]

    Improving Readability Assessment with Ordinal Log-Loss

    Lim, Ho Hung and Lee, John. Improving Readability Assessment with Ordinal Log-Loss. Proceedings of the 19th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2024). 2024

    2024

  29. [29]

    Incorporating Word-level Phonemic Decoding into Readability Assessment

    Pinney, Christine and Kennington, Casey and Pera, Maria Soledad and Landau Wright, Katherine and Fails, Jerry Alan. Incorporating Word-level Phonemic Decoding into Readability Assessment. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024

    2024

  30. [30]

    ReadMe++ : Benchmarking Multilingual Language Models for Multi-Domain Readability Assessment

    Naous, Tarek and Ryan, Michael J and Lavrouk, Anton and Chandra, Mohit and Xu, Wei. ReadMe++ : Benchmarking Multilingual Language Models for Multi-Domain Readability Assessment. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024. doi:10.18653/v1/2024.emnlp-main.682

    doi:10.18653/v1/2024.emnlp-main.682 2024

  31. [31]

    Mixture-of-Prompt-Experts for Multi-modal Semantic Understanding

    Wu, Zichen and Huang, Hsiu-Yuan and Qu, Fanyi and Wu, Yunfang. Mixture-of-Prompt-Experts for Multi-modal Semantic Understanding. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024

    2024

  32. [32]

    Going Beyond Passages: Readability Assessment for Book-level Long Texts

    Wenbiao, Li and Rui, Sun and Tianyi, Zhang and Yunfang, Wu. Going Beyond Passages: Readability Assessment for Book-level Long Texts. Proceedings of the 23rd Chinese National Conference on Computational Linguistics (Volume 1: Main Conference). 2024

    2024

  33. [33]

    Russian Journal of Linguistics , year =

    Serge Sharoff , title =. Russian Journal of Linguistics , year =

  34. [34]

    Automated text readability assessment for

    Laposhina, Antonina and Veselovskaya, Tatiana and Lebedeva, Maria and Kupreshchenko, Olga , booktitle=. Automated text readability assessment for

  35. [35]

    Peter and Fishburne, Robert P

    Kincaid, J. Peter and Fishburne, Robert P. and Rogers, Richard L. and Chissom, Benjamin S. , title =

  36. [36]

    A Language Modeling Approach to Predicting Reading Difficulty

    Collins-Thompson, Kevyn and Callan, James P. A Language Modeling Approach to Predicting Reading Difficulty. Proceedings of the Human Language Technology Conference of the North A merican Chapter of the Association for Computational Linguistics: HLT - NAACL 2004. 2004

    2004

  37. [37]

    Proceedings of the 11th Workshop on Innovative Use of NLP for Building Educational Applications (BEA) , pages =

    Text Readability Assessment for Second Language Learners , author =. Proceedings of the 11th Workshop on Innovative Use of NLP for Building Educational Applications (BEA) , pages =

  38. [38]

    Supervised and Unsupervised Neural Approaches to Text Readability

    Martinc, Matej and Pollak, Senja and Robnik- S ikonja, Marko. Supervised and Unsupervised Neural Approaches to Text Readability. Computational Linguistics. 2021. doi:10.1162/coli_a_00398

    work page doi:10.1162/coli_a_00398 2021

  39. [39]

    Anais do VII Encontro Nacional de Inteligência Artificial , location =

    Caroline Gasperin and Lucia Specia and Tiago Pereira and Sandra Aluisio , title =. Anais do VII Encontro Nacional de Inteligência Artificial , location =. 2009 , keywords =

    2009

  40. [40]

    Readability Assessment for Text Simplification

    Aluisio, Sandra and Specia, Lucia and Gasperin, Caroline and Scarton, Carolina. Readability Assessment for Text Simplification. Proceedings of the NAACL HLT 2010 Fifth Workshop on Innovative Use of NLP for Building Educational Applications. 2010

    2010

  41. [41]

    Learning to Simplify Sentences with Quasi-Synchronous Grammar and Integer Programming

    Woodsend, Kristian and Lapata, Mirella. Learning to Simplify Sentences with Quasi-Synchronous Grammar and Integer Programming. Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing. 2011

    2011

  42. [42]

    Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Student Research Workshop) , year =

    Simple, readable sub-sentences , author =. Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Student Research Workshop) , year =

  43. [43]

    Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) , year =

    Learning Simplifications for Specific Target Audiences , author =. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) , year =

  44. [44]

    Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (Student Research Workshop) , year =

    Controllable Text Simplification with Lexical Constraint Loss , author =. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (Student Research Workshop) , year =

  45. [45]

    Controlling Pre-trained Language Models for Grade-Specific Text Simplification

    Agrawal, Sweta and Carpuat, Marine. Controlling Pre-trained Language Models for Grade-Specific Text Simplification. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.790

    work page doi:10.18653/v1/2023.emnlp-main.790 2023

  46. [46]

    Mathematics , volume =

    Text Simplification to Specific Readability Levels , author =. Mathematics , volume =. 2023 , doi =

    2023

  47. [47]

    Russian Journal of Linguistics , volume=

    What neural networks know about linguistic complexity , author=. Russian Journal of Linguistics , volume=

  48. [48]

    Multiattentive Recurrent Neural Network Architecture for Multilingual Readability Assessment

    Azpiazu, Ion Madrazo and Pera, Maria Soledad. Multiattentive Recurrent Neural Network Architecture for Multilingual Readability Assessment. Transactions of the Association for Computational Linguistics. 2019. doi:10.1162/tacl_a_00278

    work page doi:10.1162/tacl_a_00278 2019

  49. [49]

    Vajjala, Sowmya and Lučić, Ivana , booktitle =

  50. [50]

    On Improving the Accuracy of Readability Classification using Insights from Second Language Acquisition

    Vajjala, Sowmya and Meurers, Detmar. On Improving the Accuracy of Readability Classification using Insights from Second Language Acquisition. Proceedings of the Seventh Workshop on Building Educational Applications Using NLP. 2012

    2012

  51. [51]

    Stenner, A. Jackson. Measuring Reading Comprehension with the Lexile Framework. Explanatory Models, Unit Standards, and Personalized Learning in Educational Measurement: Selected Papers by A. Jackson Stenner. 2023. doi:10.1007/978-981-19-3747-7_6

    work page doi:10.1007/978-981-19-3747-7_6 2023

  52. [52]

    Optimizing Statistical Machine Translation for Text Simplification

    Xu, Wei and Napoles, Courtney and Pavlick, Ellie and Chen, Quanze and Callison-Burch, Chris. Optimizing Statistical Machine Translation for Text Simplification. Transactions of the Association for Computational Linguistics. 2016. doi:10.1162/tacl_a_00107

    doi:10.1162/tacl_a_00107 2016

  53. [53]

    Creating a parallel F innish- E asy F innish dataset from news articles

    Dmitrieva, Anna and Konovalova, Aleksandra. Creating a parallel F innish- E asy F innish dataset from news articles. Proceedings of the 1st Workshop on Open Community-Driven Machine Translation. 2023

    2023

  54. [54]

    R evita: a Language-learning Platform at the Intersection of ITS and CALL

    Katinskaia, Anisia and Nouri, Javad and Yangarber, Roman. R evita: a Language-learning Platform at the Intersection of ITS and CALL. Proceedings of the Eleventh International Conference on Language Resources and Evaluation ( LREC 2018). 2018

    2018

  55. [55]

    Linguistic Constructs Represent the Domain Model in Intelligent Language Tutoring

    Katinskaia, Anisia and Hou, Jue and Vu, Anh-duc and Yangarber, Roman. Linguistic Constructs Represent the Domain Model in Intelligent Language Tutoring. Proceedings of the 17th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations. 2023. doi:10.18653/v1/2023.eacl-demo.16

    work page doi:10.18653/v1/2023.eacl-demo.16 2023

  56. [56]

    6th Workshop on NLP for CALL and 2nd Workshop on NLP for Research on Language Acquisition, at NoDaLiDa , XXXbooktitle =

    Revita: a system for language learning and supporting endangered languages , author =. 6th Workshop on NLP for CALL and 2nd Workshop on NLP for Research on Language Acquisition, at NoDaLiDa , XXXbooktitle =

  57. [57]

    Digital Cultural Heritage and Revitalization of Endangered

    Katinskaia, Anisia and Yangarber, Roman , SSSauthor =. Digital Cultural Heritage and Revitalization of Endangered. Proceedings of the 3rd Conference on Digital Humanities in the Nordic Countries , SSSbooktitle =

  58. [58]

    Modeling language learning using specialized

    Hou, Jue and Koppatz, Maximilian W and Quecedo, Jos. Modeling language learning using specialized. BEA: 14 ^

  59. [59]

    2012 , journal =

    Alex Housen, Folkert Kuiken, Ineke Vedder , title =. 2012 , journal =

    2012

  60. [60]

    Morales, Gabriel , year =

  61. [61]

    Leon Guerrero, Sibylla and Whitford, Veronica and Mesite, Laura and Luk, Gigi , title =

  62. [62]

    2024 , howpublished =

    GPT-4o System Card , author =. 2024 , howpublished =

    2024

  63. [63]

    Language Resources and Evaluation , volume =

    Democratizing Neural Machine Translation with OPUS-MT , author =. Language Resources and Evaluation , volume =. 2023 , doi =

    2023

  64. [64]

    Resources and Tools for Automatic Text Simplification: Cases of Russian and Finnish , publisher =

    Dmitrieva, Anna , year =. Resources and Tools for Automatic Text Simplification: Cases of Russian and Finnish , publisher =

  65. [65]

    S pecia L ex: A Benchmark for In-Context Specialized Lexicon Learning

    Imperial, Joseph Marvin and Tayyar Madabushi, Harish. S pecia L ex: A Benchmark for In-Context Specialized Lexicon Learning. Findings of the Association for Computational Linguistics: EMNLP 2024. 2024. doi:10.18653/v1/2024.findings-emnlp.52

    work page doi:10.18653/v1/2024.findings-emnlp.52 2024

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.