Pith. sign in

REVIEW 4 major objections 5 minor 47 references

Multiple References with Meaningful Variations Improve Literary Machine Translation

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Filtering multiple reference translations by semantic similarity, keeping only medium- and high-similarity paraphrases, improves literary machine translation over unfiltered training data.

desk verdict Worth reading for the constant-instance-count single-vs-multi result, but the headline filtering claim is confounded by a test set made of the same medium-similarity references used to build the winning training set. read the letter →

arxiv 2412.18707 v2 pith:A5V5XJFC submitted 2024-12-24 cs.CL

Si Wu , John Wieting , David A. Smith This is my paper · ORCID
classification cs.CL
keywords machinetranslationmultiplereferencesparaphrasesemanticsimilarityliterarydatafilteringfine-tuningPARAGRAM-SPlargelanguagemodels
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

Most machine-translation models are trained with a single reference translation per source, even though a sentence can be translated many ways. This paper argues that when several expert translations of the same source exist, the best recipe is not to use them all, but to filter them by semantic similarity and keep only medium- and high-similarity paraphrases. The paper reports that, across mT5-large, LLaMA-2-7B, and Opus-MT, fine-tuning on this filtered data outperforms an unfiltered multi-reference set on BLEU, COMET, and chrF++, even though the unfiltered set has more training instances. The paper also reports that when total training instances are held fixed, single-reference training with many more source texts is only marginally better than multiple-reference training with fewer source texts. If the findings generalize, data selection by similarity is a cheap way to improve literary machine translation, and multiple human translations become a valuable training resource rather than redundant text.

What carries the argument

The load-bearing mechanism is a semantic-similarity filter built on PARAGRAM-SP, a text-scoring model that returns a value from -1 to 1 for the semantic similarity of two English passages. For each source text, the paper computes $\mathrm{sim}_p$ as the average pairwise PARAGRAM-SP score over all its reference translations, and thresholds at 0.45 and 0.85 define Low, Medium, and High similarity. This filter is what turns the raw Par3 translations into the Medium, Medium+Low, and Medium+High training sets, so the paper's central intervention is training on similarity-filtered references rather than on all references together.

What would settle it

Construct held-out test sets whose references are low-similarity, high-similarity, or a mixture matching the unfiltered distribution, and re-run the Medium+High versus Unfiltered comparison on them; if the Medium+High advantage disappears or reverses, the paper's recommendation is an artifact of test-set construction.

Watch

Extended reading notes

Core claim

The central claim is that the similarity structure inside a multi-reference training set matters as much as its size. Using PARAGRAM-SP to score pairwise similarity among English translations of each source text, the paper defines $\mathrm{sim}_p$ as the average over all reference pairs and partitions references into Low (below 0.45), Medium (0.45–0.85), and High (above 0.85). Across mT5-large, LLaMA-2-7B, and Opus-MT, fine-tuning on the Medium+High filtered set yields the best overall literary MT performance, with the two LLMs showing the clearest gains over Unfiltered: 0.3–0.5 BLEU, 0.1–0.9 COMET, and 0.17–0.32 chrF++. Adding Low-similarity references to Medium makes performance worse than Unfiltered, which the paper attributes to confusing or misaligned paraphrases. Separate from filtering, the paper finds that holding #INSTANCE constant, Single-reference training with at least twice the source texts only marginally outperforms Medium multi-reference training.

Load-bearing premise

The load-bearing premise is that the test set's medium-similarity references are the right evaluation condition, since the best training set is selected to contain that same similarity band and the reported gains could vanish if test references were low- or high-similarity.

Editorial extensions

If this is right

  • Fine-tuning on Medium+High references gives the best literary MT performance across the three models and, for mT5-large and LLaMA-2-7B, beats an Unfiltered training set with more instances by 0.3–0.5 BLEU, 0.1–0.9 COMET, and 0.17–0.32 chrF++.
  • At a fixed total instance count, Single-reference training with at least twice as many source texts differs from Medium multi-reference training by at most 0.7 BLEU, 2.5 COMET, and 1.98 chrF++, so multiple references can nearly substitute for additional source texts.
  • Adding Low-similarity references to a Medium training set makes performance worse than Unfiltered, indicating that noisy or misaligned paraphrases hurt rather than help.
  • The filtering benefit is clearer for the two LLMs than for the smaller Opus-MT model, suggesting that model capacity shapes how much varied references can be exploited.

Reading between the lines

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

  • A natural extension is to apply the same $\mathrm{sim}_p$ filter to synthetic paraphrases generated by LLMs; if the threshold transfers, it would give a low-cost cleaning step for domains where multiple human translations do not exist.
  • The near-parity between single- and multi-reference training at fixed instance count suggests that when expert multi-references already exist, they are almost as valuable as new source texts; a cost comparison of annotation effort could make this concrete.
  • Because the test set contains only medium-similarity references, the optimal threshold is likely tuned to that evaluation condition; a deployment evaluating on low- or high-similarity references may need to re-select the band.
  • The finding that Low-similarity references actively hurt performance implies that simply collecting more translations is not always safer than collecting fewer, which may extend to other multilingual data-selection tasks.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper investigates how to best use multiple human reference translations when fine-tuning literary machine translation models. Using the Par3 dataset, it scores the semantic similarity between English references with PARAGRAM-SP and partitions source texts into Low, Medium, and High similarity groups. It constructs Single, Medium, Unfiltered, Medium+Low, and Medium+High training sets, and fine-tunes mT5-large, LLaMA-2-7B, and Opus-MT. Two main findings are reported: (i) holding total training instances constant, single-reference training with more source texts is only marginally better than multiple-reference training with half the sources; (ii) filtering to Medium+High similarity references outperforms an unfiltered dataset, with gains of 0.3–0.5 BLEU, 0.1–0.9 COMET, and 0.17–0.32 chrF++ for the LLMs. The paper provides public code and uses paired bootstrap significance tests.

Significance. If the findings hold, they offer practical guidance for multi-reference MT training and demonstrate the value of semantic-similarity filtering. The study has notable strengths: it uses real expert literary translations, controls for total training instances in the single-versus-multiple comparison, evaluates three diverse model architectures, and applies bootstrap significance testing. The code is publicly available, which supports reproducibility. However, the central recommendation about filtering low-similarity references is currently validated only against a test set that shares the same similarity filter, which limits its generalizability until that confound is addressed.

major comments (4)
  1. [Section 4 and Section 6.2] The evaluation set is constructed to contain only medium-similarity references, which is the same property used to build the Medium and Medium+High training sets. Section 4 states: 'The test dataset is used in all experiments, and it has multiple human references per source text for evaluation. It is composed of medium semantic similarity only.' Consequently, the comparison between Unfiltered and Medium+High training in Table 3 confounds the effect of reference similarity with train/eval distribution matching: a model trained on Medium+High sees training targets that resemble the test references, while Unfiltered training includes low- and high-similarity targets that may shift output style away from the test distribution. The observed gains are therefore consistent with an artifact of test-set construction. The paper should either evaluate on an unfiltered test set covering all similarity levels, or report results separately for low- and high-similarity test references. The Limitations section does not mention this confound.
  2. [Section 6.2, Table 3] The claim that 'Medium+High (#SOURCE = 66k, 68k) outperforms Unfiltered across different MT metrics and models (p-value < 0.01)' is not supported by the data in Table 3. For Opus-MT, BLEU at #SRC=66k is tied at 25.3 (marked †, not significant) and at #SRC=68k is lower than Unfiltered (25.1 vs 25.3, marked †). The significant improvements are limited to mT5 and LLaMA, and for Opus only on chrF++ at #SRC=66k. The abstract's phrasing 'when fine-tuning an LLM' is appropriate; the broader 'MT model' phrasing in Section 6.2 and the bullet points in the Introduction overstate the results. Please restrict the claim accordingly or provide justification for excluding Opus-MT's BLEU scores.
  3. [Section 6.1, Table 2] The text reports that 'Single on average only marginally outperforms Medium ... ranging 0.0–0.7 in BLEU.' However, Table 2 shows LLaMA-2-7B at #INSTANCE=77k has a BLEU difference of 1.2 (29.4 vs 28.2), and Opus-MT at #INSTANCE=58k has a difference of 0.8 (23.9 vs 23.1). The stated range is therefore incorrect and should be 0.0–1.2. While the qualitative conclusion of marginal superiority may still hold, the misreported range should be corrected.
  4. [Section 3 and Section 4] The threshold for 'meaningful variations' (sim_p in [0.45, 0.85]) is selected by manual inspection, and the same threshold is used to define both the training conditions and the test set. No sensitivity analysis is provided for this choice. Because the main recommendation is to filter by this threshold, the paper should demonstrate that the results are stable across nearby threshold values or at least discuss how the threshold was validated. This is particularly important given that the similarity scorer (PARAGRAM-SP) is co-developed by one of the authors, which increases the burden of showing the results are not an artifact of the specific scoring choice.
minor comments (5)
  1. [Section 4] The description of the Single dataset should state whether its references are drawn from all similarity levels in Par3 or only from a particular subset. This affects the interpretation of the Single vs. Medium comparison in Table 2.
  2. [Section 4] The paper does not report the size or language breakdown of the test set; this information should be added for reproducibility.
  3. [Section 6.5, Table 4] The Medium+1.0 condition is constructed by repeating references from other categories, which may not be equivalent to naturally occurring identical paraphrases; this limitation should be acknowledged.
  4. [Section 5.3] There is a typo in Section 5.3: 'randolmized' should be 'randomized'.
  5. [Section 4] The notation #SOURCE and #INSTANCE is used throughout; an explicit definition table in Section 4 would improve readability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main comparisons are empirical and self-contained; the medium-only test set is a limitation, not a construction-level circularity.

full rationale

No step in the paper reduces by construction to its own inputs. The similarity thresholds (medium = [0.45, 0.85]) were fixed by manual inspection before the fine-tuning experiments, and the central result in Table 3 is an empirical outcome: the Unfiltered dataset has more training instances (181k vs 159k for Medium+High), so it could plausibly have won; the reported BLEU/COMET/chrF++ improvements are not entailed by the filter definitions. The self-citation to PARAGRAM-SP (Wieting et al., 2022) introduces a measuring instrument rather than a uniqueness theorem or a smuggled ansatz, and the paper's conclusions do not depend on asserting the scorer's effectiveness as a matter of self-citation alone. The closest thing to a circular concern is Section 4's statement that the test dataset is 'composed of medium semantic similarity only,' which makes the evaluation distribution similar to the favored Medium+High training distribution and could explain part of the advantage through train/eval distribution matching. However, this is an experimental design limitation, not a case where the prediction is equivalent to its inputs: the metric values are not forced by the test-set construction, and the comparison retains independent empirical content.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claims rest on the assumption that PARAGRAM-SP similarity bands capture meaningful translation variation, and on a test set that only uses medium-similarity references. The thresholds 0.45 and 0.85 are free parameters set by manual inspection. No new entities are introduced.

free parameters (4)
  • Lower threshold for Medium semantic similarity = 0.45
    Chosen by manual inspection of the sim_p distribution to define 'meaningful variations'; no independent derivation.
  • Upper threshold for Medium semantic similarity = 0.85
    Chosen by manual inspection; the paper says paraphrases in this range are 'different enough to be interesting, yet similar enough'.
  • Test set composition = medium similarity only, [0.45, 0.85)
    Evaluation set includes only medium-similarity references, which may favor training sets built from the same filter.
  • Incremental addition step for Low/High references = 10%
    The paper constructs Medium+Low and Medium+High by adding 10% of Low or High references at each step until exhausted; the step size is arbitrary.
assumptions (4)
  • domain assumption PARAGRAM-SP scores are a valid measure of semantic similarity between literary translations for this purpose.
    Used in Section 3 to classify all references; the paper verifies the qualitative bands only by manual inspection.
  • ad hoc to paper The manually selected sim_p range [0.45, 0.85] corresponds to 'meaningful variations' relevant for MT training.
    Defined after inspecting the distribution; not derived from translation theory or prior benchmarks.
  • domain assumption Fine-tuning on multiple references as separate training instances is an appropriate way to exploit multiple references.
    Standard MT practice, but the paper does not compare with other multi-reference training objectives such as lattices or weighted references.
  • domain assumption Paragraph-level translation units improve literary cohesion and are suitable for this study.
    The paper follows Voigt and Jurafsky (2012) to use paragraphs rather than sentences; this choice affects what variations are possible.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multiple References with Meaningful Variations Improve Literary Machine Translation." pith.science (2026). https://pith.science/paper/A5V5XJFC

@misc{pith2026241218707,
  author       = {Pith},
  title        = {Pith review of: Multiple References with Meaningful Variations Improve Literary Machine Translation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A5V5XJFC}},
  note         = {Machine review of arXiv:2412.18707}
}
read the original abstract

While a source sentence can be translated in many ways, most machine translation (MT) models are trained with only a single reference. Previous work has shown that using synthetic paraphrases can improve MT. This paper investigates best practices for employing multiple references by analyzing the semantic similarity among different English translations of world literature in the Par3 dataset. We classify the semantic similarity between paraphrases into three levels: low, medium, and high, and fine-tune three different models (mT5-large, LLaMA-2-7B, and Opus-MT) for literary MT tasks. Across different models, holding the total training instances constant, single-reference but more source texts only marginally outperforms multiple-reference with half of the source texts. Moreover, when fine-tuning an LLM, using paraphrases with medium and high semantic similarity outperforms an unfiltered dataset, with improvements in BLEU (0.3-0.5), COMET (0.1-0.9), and chrF++ (0.17-0.32). Our code is publicly available on GitHub.

Figures

Figures reproduced from arXiv: 2412.18707 by the authors.

Figure 1
Figure 1. An example of meaningful variations between [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The distribution of sim_p scores for the source [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Performance difference between different languages under different MT metrics when fine-tuning with [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 14 canonical work pages

  1. [1]

    URL: " 'urlintro :=

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

  2. [2]

    write newline

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

  3. [3]

    Colin Bannard and Chris Callison-Burch. 2005. https://doi.org/10.3115/1219840.1219914 Paraphrasing with bilingual parallel corpora . In Proceedings of the 43rd Annual Meeting of the Association for Computational Linguistics ( ACL ' 05) , pages 597--604, Ann Arbor, Michigan. Association for Computational Linguistics

    arXiv 2005

  4. [4]

    Rahul Bhagat and Eduard Hovy. 2013. https://doi.org/10.1162/COLI_a_00166 S quibs: What is a paraphrase? Computational Linguistics, 39(3):463--472

    doi:10.1162/coli_a_00166 2013

  5. [5]

    Chris Callison-Burch, Philipp Koehn, and Miles Osborne. 2006. https://aclanthology.org/N06-1003 Improved statistical machine translation using paraphrases . In Proceedings of the Human Language Technology Conference of the NAACL , Main Conference , pages 17--24, New York City, USA. Association for Computational Linguistics

    work page 2006

  6. [6]

    J.C. Catford. 1965. A Linguistic Theory of Translation: An Essay in Applied Linguistics. Oxford University Press

    work page 1965

  7. [7]

    Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzm \'a n, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2020. https://doi.org/10.18653/v1/2020.acl-main.747 Unsupervised cross-lingual representation learning at scale . In Proceedings of the 58th Annual Meeting of the Association for Comp...

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

  8. [8]

    Zachary Horvitz, Ajay Patel, Chris Callison-Burch, Zhou Yu, and Kathleen McKeown. 2024. http://arxiv.org/abs/2308.15459 Paraguide: Guided diffusion paraphrasers for plug-and-play textual style transfer

    arXiv 2024

Show all 47 references
  1. [9]

    Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. https://openreview.net/forum?id=nZeVKeeFYf9 Lo RA : Low-rank adaptation of large language models . In International Conference on Learning Representations

    2022

  2. [10]

    Lasal Jayawardena and Prasan Yapa. 2024 a . https://doi.org/10.5121/csit.2024.140418 Parafusion: A large-scale llm-driven english paraphrase dataset infused with high-quality lexical and syntactic diversity . In Artificial Intelligence and Big Data, AIBD. Academy & Industry Re...

    2024

  3. [11]

    Lasal Jayawardena and Prasan Yapa. 2024 b . https://doi.org/10.1109/icacs60934.2024.10473289 Parameter efficient diverse paraphrase generation using sequence-level knowledge distillation . In 2024 5th International Conference on Advancements in Computational Sciences (ICACS). IEEE

    2024

  4. [12]

    Marcin Junczys-Dowmunt, Roman Grundkiewicz, Tomasz Dwojak, Hieu Hoang, Kenneth Heafield, Tom Neckermann, Frank Seide, Ulrich Germann, Alham Fikri Aji, Nikolay Bogoychev, Andr \'e F. T. Martins, and Alexandra Birch. 2018. https://doi.org/10.18653/v1/P18-4020 M arian: Fast neura...

    2018 doi

  5. [13]

    Huda Khayrallah and Philipp Koehn. 2018. https://doi.org/10.18653/v1/W18-2709 On the impact of various types of noise on neural machine translation . In Proceedings of the 2nd Workshop on Neural Machine Translation and Generation, pages 74--83, Melbourne, Australia. Associatio...

    2018 doi

  6. [14]

    Huda Khayrallah, Brian Thompson, Matt Post, and Philipp Koehn. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.7 Simulated multiple reference training improves low-resource machine translation . In Proceedings of the 2020 Conference on Empirical Methods in Natural Language P...

    2020 doi

  7. [15]

    Philipp Koehn. 2004. https://aclanthology.org/W04-3250 Statistical significance tests for machine translation evaluation . In Proceedings of the 2004 Conference on Empirical Methods in Natural Language Processing, pages 388--395, Barcelona, Spain. Association for Computational...

    2004

  8. [16]

    Andr\'e Lefevere. 1992. Translating literature : practice and theory in a comparative literature context. New York :Modern Language Association of America

    1992

  9. [17]

    Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Veselin Stoyanov, and Luke Zettlemoyer. 2020. https://doi.org/10.18653/v1/2020.acl-main.703 BART : Denoising sequence-to-sequence pre-training for natural language generation, translatio...

    2020 doi

  10. [18]

    Nitin Madnani, Necip Fazil Ayan, Philip Resnik, and Bonnie Dorr. 2007. https://aclanthology.org/W07-0716 Using paraphrases for parameter tuning in statistical machine translation . In Proceedings of the Second Workshop on Statistical Machine Translation, pages 120--127, Prague...

    2007

  11. [19]

    Dorr, and Richard Schwartz

    Nitin Madnani, Philip Resnik, Bonnie J. Dorr, and Richard Schwartz. 2008. https://aclanthology.org/2008.amta-papers.13 Are multiple reference translations necessary? investigating the value of paraphrased reference translations in parameter optimization . In Proceedings of the...

    2008

  12. [20]

    Jean Maillard, Cynthia Gao, Elahe Kalbassi, Kaushik Ram Sadagopan, Vedanuj Goswami, Philipp Koehn, Angela Fan, and Francisco Guzman. 2023. https://doi.org/10.18653/v1/2023.acl-long.154 Small data, big impact: Leveraging minimal data for effective machine translation . In Proce...

    2023 doi

  13. [21]

    Kathleen R. McKeown. 1983. https://aclanthology.org/J83-1001 Paraphrasing questions using given and new information . American Journal of Computational Linguistics, 9(1):1--10

    1983

  14. [22]

    Abhika Mishra, Akari Asai, Vidhisha Balachandran, Yizhong Wang, Graham Neubig, Yulia Tsvetkov, and Hannaneh Hajishirzi. 2024. http://arxiv.org/abs/2401.06855 Fine-grained hallucination detection and editing for language models

    2024 arXiv

  15. [23]

    E.A. Nida. 1964. Toward a Science of Translating: With Special Reference to Principles and Procedures Involved in Bible Translating (Second edition). Brill

    1964

  16. [24]

    NLLB Team , Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia Gonzalez, Pran...

    2022 arXiv

  17. [25]

    Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. https://doi.org/10.3115/1073083.1073135 B leu: a method for automatic evaluation of machine translation . In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311--3...

    2002

  18. [26]

    Maja Popovi \'c . 2017. https://doi.org/10.18653/v1/W17-4770 chr F ++: words helping character n-grams . In Proceedings of the Second Conference on Machine Translation, pages 612--618, Copenhagen, Denmark. Association for Computational Linguistics

    2017 doi

  19. [27]

    Ricardo Rei, Jos \'e G. C. de Souza, Duarte Alves, Chrysoula Zerva, Ana C Farinha, Taisiya Glushkova, Alon Lavie, Luisa Coheur, and Andr \'e F. T. Martins. 2022. https://aclanthology.org/2022.wmt-1.52 COMET -22: Unbabel- IST 2022 submission for the metrics shared task . In Pro...

    2022

  20. [28]

    Ricardo Rei, Craig Stewart, Ana C Farinha, and Alon Lavie. 2020. https://doi.org/10.18653/v1/2020.emnlp-main.213 COMET : A neural framework for MT evaluation . In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2685--2702, ...

    2020 doi

  21. [29]

    Shiqi Shen, Yong Cheng, Zhongjun He, Wei He, Hua Wu, Maosong Sun, and Yang Liu. 2016. https://doi.org/10.18653/v1/P16-1159 Minimum risk training for neural machine translation . In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume ...

    2016 doi

  22. [30]

    Smith, and Luke Zettlemoyer

    Gabriel Stanovsky, Noah A. Smith, and Luke Zettlemoyer. 2019. https://doi.org/10.18653/v1/P19-1164 Evaluating gender bias in machine translation . In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1679--1684, Florence, Italy. Ass...

    2019 doi

  23. [31]

    Dan Su, Xiaoguang Li, Jindi Zhang, Lifeng Shang, Xin Jiang, Qun Liu, and Pascale Fung. 2022. http://arxiv.org/abs/2203.00343 Read before generate! faithful long form question answering with machine reading

    2022 arXiv

  24. [32]

    Katherine Thai, Marzena Karpinska, Kalpesh Krishna, Bill Ray, Moira Inghilleri, John Wieting, and Mohit Iyyer. 2022. https://doi.org/10.18653/v1/2022.emnlp-main.672 Exploring document-level literary machine translation with parallel paragraphs from world literature . In Procee...

    2022 doi

  25. [33]

    o rg Tiedemann, Mikko Aulamo, Daria Bakshandaeva, Michele Boggia, Stig-Arne Gr \

    J \"o rg Tiedemann, Mikko Aulamo, Daria Bakshandaeva, Michele Boggia, Stig-Arne Gr \"o nroos, Tommi Nieminen, Alessandro Raganato\, Yves Scherrer, Raul Vazquez, and Sami Virpioja. 2023. https://doi.org/10.1007/s10579-023-09704-w Democratizing neural machine translation with OP...

    2023 doi

  26. [34]

    J \"o rg Tiedemann and Santhosh Thottingal. 2020. OPUS-MT — B uilding open translation services for the W orld. In Proceedings of the 22nd Annual Conferenec of the European Association for Machine Translation (EAMT), Lisbon, Portugal

    2020

  27. [35]

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, W...

    2023 arXiv

  28. [36]

    Rob Voigt and Dan Jurafsky. 2012. https://aclanthology.org/W12-2503 Towards a literary machine translation: The role of referential cohesion . In Proceedings of the NAACL - HLT 2012 Workshop on Computational Linguistics for Literature , pages 18--25, Montr \'e al, Canada. Asso...

    2012

  29. [37]

    Jan Philip Wahle, Bela Gipp, and Terry Ruas. 2023. https://doi.org/10.18653/v1/2023.emnlp-main.746 Paraphrase types for generation and detection . In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 12148--12164, Singapore. Associat...

    2023 doi

  30. [38]

    Anna Wegmann, Tijs van den Broek, and Dong Nguyen. 2024. http://arxiv.org/abs/2404.06670 What's mine becomes yours: Defining, annotating and detecting context-dependent paraphrases in news interview dialogs

    2024 arXiv

  31. [39]

    John Wieting, Taylor Berg-Kirkpatrick, Kevin Gimpel, and Graham Neubig. 2019 a . https://doi.org/10.18653/v1/P19-1427 Beyond BLEU : Training neural machine translation with semantic similarity . In Proceedings of the 57th Annual Meeting of the Association for Computational Lin...

    2019 doi

  32. [40]

    John Wieting, Kevin Gimpel, Graham Neubig, and Taylor Berg-Kirkpatrick. 2019 b . https://doi.org/10.18653/v1/P19-1453 Simple and effective paraphrastic similarity from parallel translations . In Proceedings of the 57th Annual Meeting of the Association for Computational Lingui...

    2019 doi

  33. [41]

    John Wieting, Kevin Gimpel, Graham Neubig, and Taylor Berg-kirkpatrick. 2022. https://doi.org/10.18653/v1/2022.emnlp-demos.38 Paraphrastic representations at scale . In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstration...

    2022 doi

  34. [42]

    Sander Wubben, Antal van den Bosch, and Emiel Krahmer. 2010. https://aclanthology.org/W10-4223 Paraphrase generation as monolingual translation: Data and evaluation . In Proceedings of the 6th International Natural Language Generation Conference. Association for Computational ...

    2010

  35. [43]

    Haoran Xu, Young Jin Kim, Amr Sharaf, and Hany Hassan Awadalla. 2024. https://openreview.net/forum?id=farT6XXntP A paradigm shift in machine translation: Boosting translation performance of large language models . In The Twelfth International Conference on Learning Representations

    2024

  36. [44]

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

    2021 doi

  37. [45]

    Xuan Zhang, Navid Rajabi, Kevin Duh, and Philipp Koehn. 2023. https://doi.org/10.18653/v1/2023.wmt-1.43 Machine translation with large language models: Prompting, few-shot learning, and fine-tuning with QL o RA . In Proceedings of the Eighth Conference on Machine Translation, ...

    2023 doi

  38. [46]

    Renjie Zheng, Mingbo Ma, and Liang Huang. 2018. https://doi.org/10.18653/v1/D18-1357 Multi-reference training with pseudo-references for neural translation and text generation . In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 31...

    2018 doi

  39. [47]

    Vil \'e m Zouhar, Shuoyang Ding, Anna Currey, Tatyana Badeka, Jenyuan Wang, and Brian Thompson. 2024. https://doi.org/10.18653/v1/2024.acl-short.45 Fine-tuned machine translation metrics struggle in unseen domains . In Proceedings of the 62nd Annual Meeting of the Association ...

    2024 doi

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.