TopXGen: Topic-Diverse Parallel Data Generation for Low-Resource Machine Translation

Armel Zebaze; Beno\^it Sagot; Rachel Bawden

REVIEW 4 major objections 6 minor 7 references

TopXGen: Topic-Diverse Parallel Data Generation for Low-Resource Machine Translation

T0 review · 4 major / 6 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 An LLM that translates poorly into low-resource languages can still write fluent text in them, and backtranslating that text into English yields parallel data that makes smaller models translate well into those languages.

desk verdict A well-controlled empirical study showing topic-guided LLM generation plus backtranslation improves low-resource MT, with the main caveat that target-side quality is only indirectly verified. read the letter →

arxiv 2508.08680 v1 pith:TYP47B4K submitted 2025-08-12 cs.CL

Armel Zebaze , Benoît Sagot , Rachel Bawden This is my paper · ORCID

classification cs.CL

keywords low-resourcemachinetranslationsyntheticparalleldatatopic-guidedgenerationbacktranslationin-contextlearningfine-tuningmultilingualLLMsFLORES-200

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

The paper tries to establish that the bottleneck for low-resource machine translation is not the translator but the parallel data: a strong multilingual LLM, asked to write paragraphs about arbitrary topics directly in ten low-resource languages, produces target-side text that can be backtranslated into English to form synthetic sentence-level parallel data. Fine-tuning small decoder models on this data moves them from near-random translation to scores that rival or beat far larger instruction-tuned models, and retrieving examples from the synthetic pool for in-context learning beats retrieving from the in-domain human-written development set on the reference-free quality metric. This matters because it suggests that high-quality low-resource MT training data can be manufactured cheaply, without parallel corpora or expert translators, wherever the generator has basic competence in the target language.

What carries the argument

The carrier of the argument is a three-part generation prompt: a randomly drawn topic from a 67,573-entry Wikipedia topic list (for content diversity), 240 seed paragraphs from XQuAD in high-resource languages (for expected length and format via cross-lingual in-context learning), and FLORES-dev seed sentences in the target low-resource language (for script, register, and sentence shape). Paragraphs written by Gemma-3-27B-It are sentence-split, run through fastText language identification, filtered for overlap against FLORES, NTREX, TICO-19, and the XQuAD seeds, and then backtranslated into English with NLLB-200-3.3B. Topic diversity and a temperature of 1.0 are the tuning knobs that produce the strongest student models, and the same synthetic pool serves both fine-tuning and retrieval-based few-shot prompting.

What would settle it

Run TOPXGEN on a language where the generator's target-side writing is known to be weak, or ask native speakers to rate a sample of generated sentences for fluency and correctness, and check whether the fine-tuned student still beats both a zero-shot strong model and a forward-translation baseline; if fluency breaks, the method should collapse toward that baseline.

Watch

Extended reading notes

Core claim

On its own terms, the discovery is that target-side quality is the lever: instead of translating English text into a low-resource language (forward translation) or mining scarce monolingual corpora, TOPXGEN prompts an LLM to produce original paragraphs in the target language, guided by a random topic, by high-resource-language seed paragraphs, and by a handful of target-language seed sentences. These paragraphs are split into sentences, filtered for language, deduplicated, and backtranslated into English with NLLB-200-3.3B. The resulting 1.05M English-to-low-resource-language sentence pairs, about 50k to 150k per language, let a fine-tuned LLaMA-3-8B outperform Gemma-2-27B-It and LLaMA-3.1-70B-It in the tested English-to-LRL directions and come close to the Gemma-3-27B-It generator itself, while a 5-shot in-context learner retrieving from TOPXGEN beats the same retrieval from the FLORES dev set on MetricX-24. The paper also shows the pipeline beats Self-Instruct and KNN-Instruct under identical data budgets, and that data scale plus topic diversity, not just quality, is what drives the gains.

Load-bearing premise

The pipeline assumes the generator LLM can already write fluent, grammatical, script-correct text in each target low-resource language, and that fluency is verified only indirectly by language identification, quality estimates, and downstream translation scores rather than by human judgement.

Editorial extensions

If this is right

Fine-tuning an 8B model on TOPXGEN data gives translations into ten low-resource languages that match or beat instruction-tuned models several times its size, with about three hours of training on a single H100.
Retrieving five in-context examples from the synthetic pool outperforms retrieving from the in-domain FLORES dev set on MetricX-24, so the generated data can replace human-written development sets as an example bank.
The generator can be swapped: using GPT-4o-mini to generate the target text yields stronger students than Gemma-3-27B-It, and using either the generator or the fine-tuned student as the backtranslator works almost as well as NLLB.
More topics mean better student models, so the pipeline's value should grow with topic-list coverage rather than with sentence count alone.
Scaling from roughly 900 professionally translated sentences to the roughly 100-times-larger synthetic set outweighs the quality edge of professional data on the tested directions.

Reading between the lines

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

An extension the paper does not test is pushing TOPXGEN into languages where the generator is clearly weak, since the stated limitation predicts the method will degrade toward forward translation exactly when target-side fluency fails.
The ablation that swaps back-translators (NLLB vs generator vs student) suggests the English side is not the bottleneck; a testable consequence is that investing in better low-resource-language generation, not better back-translation, should drive further gains.
Because more topics helped even when the topic list was cut from 67,573 to 509, the topic list itself acts as a diversity budget; curating topics per language or per domain could be a cheap way to steer the data.
A head-to-head against human-written monolingual corpora of equal size would isolate whether topic-diverse generation or backtranslation-into-a-high-resource-language is the active ingredient.

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.

Desk Editor's Note

read the letter

TopXGen is worth a serious referee. The core idea is simple and sensible: instead of needing existing LRL monolingual corpora for backtranslation, prompt a multilingual LLM to generate topic-diverse LRL text using seed sentences, then backtranslate to English with NLLB. This extends Marie and Fujita's idea of LLM-synthesized monolingual data, but removes the need for domain fine-tuning and works directly in ten LRLs. That is a practically useful advance for low-resource MT.

What the paper does well: the experimental setup is careful. Evaluation is on FLORES, NTREX, and TICO with BLEU and MetricX, paired bootstrap significance tests, contamination filtering toward evaluation sets, and several ablations (generator, backtranslator, temperature, number of topics, self-improvement). The comparison against Self-Instruct and KNN-Instruct under the same budget is clean and consistently favors TopXGen. The ICL result, where retrieval from TopXGen beats retrieval from the in-domain FLORES dev set on MetricX, is a strong point. Code and outputs are released.

The soft spots are real but not fatal. The load-bearing premise is that Gemma-3-27B-It writes fluent, script-correct Hausa, Igbo, Xhosa, etc. The authors never get human judgments of the generated LRL text. They rely on fastText LID, MetricX QE, LLM topic-alignment, and a Google Translate cross-check. None of these directly verify naturalness. The B.8 cross-check only shows that two MT systems produce similar English from the generated LRL, which is weak evidence about target-side quality. The temperature ablation shows quality collapses above T=1.2, so the pipeline is fragile to generator choice. However, the authors explicitly acknowledge the generator requirement in Limitations, and the downstream MT gains are the real payoff. If the generated text were garbage, you would not expect the fine-tuned LLaMA models to beat strong baselines and approach the generator's own score. So the central comparative claim - TopXGen beats Self-Instruct and KNN-Instruct under identical budgets - holds up.

One smaller complaint: the abstract and conclusion say the datasets are 'comparable to professionally written text', which overshoots. Figure 3 shows TopXGen does not consistently beat small professional datasets at equal scale; it wins because it can be 100x larger. That is a legitimate advantage, but the claim should be phrased as 'scale compensates for quality differences', not 'comparable quality'.

The citation pattern is solid. They properly credit Marie and Fujita as the closest prior work, and they are not hiding the relationship.

Who should read it: anyone working on data synthesis for low-resource MT, and anyone evaluating synthetic data methods. It deserves peer review. I would accept it as a serious submission, with a request for either human evaluation of generated target text or a narrowing of the quality claim.

Referee Report

4 major / 6 minor

Summary. The paper presents TopXGen, a pipeline that prompts a multilingual LLM (Gemma-3-27B-It) to generate topic-guided paragraphs directly in ten low-resource languages, then backtranslates them into English with NLLB-200-3.3B to obtain synthetic parallel sentence pairs. The authors evaluate the resulting data by fine-tuning LLaMA-2-7B and LLaMA-3-8B and by performing 5-shot in-context learning with LLaMA-3.1-8B-It, using FLORES-200, NTREX-128, and TICO-19 with BLEU, chrF++, MetricX-24, and XCOMET-XL. They compare against Self-Instruct and KNN-Instruct data under the same training budget and against small professionally translated datasets, and they run ablations on the generator, topic-list size, temperature, back-translator, and iterative self-improvement. The central claim is that TopXGen data improves translation into low-resource languages, with unidirectional fine-tuned models approaching or exceeding the generator's zero-shot quality, and that retrieval from TopXGen gives better MetricX-24 scores than retrieval from the in-domain FLORES dev set.

Significance. If the results hold, the paper makes a practical contribution: a scalable and relatively low-cost recipe for producing synthetic parallel data for low-resource machine translation without relying on existing target-side monolingual corpora. The evaluation is externally grounded in human-translated references, includes contamination filtering and significance testing, reports multiple metrics and benchmarks, and provides valuable ablations. The code and outputs are released, which supports reproducibility. The main caveat is that the quality of the generated LRL text is inferred from automatic and indirect signals rather than from human judgments, and the headline comparison to prior data-generation methods covers only two languages and one base model.

major comments (4)

[Section 3 and Limitations] The central premise that Gemma-3-27B-It produces fluent, grammatical, script-consistent text in all ten LRLs is not directly validated. The evidence is indirect: fastText LID checks only the language label; MetricX-24 QE is a learned estimator; the LLM topic-alignment judgments in Table 15 check topic adherence, not target-side naturalness; and the Google Translate cross-check in Table 17 checks semantic consistency with the backtranslated English. The temperature ablation in Section 6 shows that generation degrades sharply above T=1.2, confirming that output quality is fragile and generator-dependent. Because the abstract and introduction claim 'high quality and natural-sounding' target text, I request a human (or native-speaker) evaluation of a sample of generated sentences per language, together with the fastText LID retention rates, so that the data-quality premise is supported directly rather than only by downstream MT gains.
[Section 5.3 (Comparison to Self-Instruct and KNN-Instruct)] The claim that TopXGen 'consistently outperforms' Self-Instruct and KNN-Instruct is supported only for two languages (Sundanese and Somali) and one base model (LLaMA-2-7B) at 20K sentence pairs. This is a narrow basis for the general statement in the abstract. I recommend either adding at least one additional language/base-model setting for the comparison, or explicitly qualifying the claim to the tested directions.
[Section 5.1 (Fine-tuning results)] The statement that unidirectional LLaMA-3-8B fine-tuning 'outperforms Gemma-2-27B-It and LLaMA-3.1-70B-It' is not supported by Table 2 for every direction and metric: for example, Basque BLEU is 25.64 for the fine-tuned model versus 26.06 for LLaMA-3.1-70B-It, and Swahili BLEU is 33.49 versus 35.99 for Gemma-2-27B-It, with MetricX also worse in the Swahili case. Please report averages across the ten directions or otherwise qualify the claim by metric and language so that the conclusion reflects the actual pattern of results.
[Section 3.1 and Related Work] The paper does not compare TopXGen against standard backtranslation applied to existing target-side monolingual corpora (for example, CC-100, OSCAR, or language-specific news corpora). Since TopXGen is precisely a method for creating target-side text to backtranslate, such a baseline is needed to determine whether the gains come from topic-guided generation and diversity or simply from having additional target-side data. Please add this comparison or clearly justify its absence for the ten selected languages.

minor comments (6)

[Section 4.1] The phrase 'as they cannot follows instructions' should read 'as they cannot follow instructions'.
[Appendix B.4] The text says 'We report the results obtained on NTREX-19 in Table 11,' but the benchmark is NTREX-128; please correct the identifier.
[Appendix B.8] The sentence 'we observe that Gemma-3-27B-It's generations in share lexical overlap' appears to be missing a language specification; please rephrase.
[Limitations] The phrase 'monolingual data scrapped from the internet' should be 'scraped from the internet'.
[Tables 3, 5, and 6] The tables rely on underline markers to indicate statistical equivalence to the best score, but these markers are not visible in the submitted text; please ensure they are rendered in the final version.
[Throughout] The paper alternates between 'SMOLSENT' and 'SMOL'; please unify the notation.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: headline gains are measured on external human references; the generator-fluency premise is an acknowledged, empirically tested limitation, not a built-in assumption.

full rationale

The central evaluation chain is not circular. TOPXGEN constructs synthetic parallel data by prompting Gemma-3-27B-It to write LRL paragraphs and backtranslating them with NLLB-200-3.3B, and every headline result is scored against human-translated FLORES-200 devtest, NTREX-128, and TICO-19 references that are not used to build the training data. The pipeline removes 10-gram overlaps with these benchmarks, so the fine-tuned student cannot simply memorize test sentences. The load-bearing premise—that the generator can produce fluent, script-consistent text in the ten LRLs—is explicitly acknowledged in the Limitations section and is probed by generator swaps, temperature ablations, fastText filtering, and external downstream MT metrics; it is not assumed as the proof of the method. The topic-alignment check in Table 15 is partly self-referential because the generator judges its own output, but an independent judge (Llama-4-Scout) is also used and this check is not the basis of the central claim. Self-citations (Zebaze et al., 2025a,b) appear in related work and as a retrieval/example-selection choice, but they are not used to justify TOPXGEN's effectiveness and are not load-bearing. No equation or fitted parameter is reused as a prediction, and no specific reduction from input to output can be exhibited. The mild scent noted by a skeptical reader is a data-quality verification gap, not a circular derivation.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The pipeline's success depends on a small set of domain assumptions: the generator writes good LRL text, the backtranslator preserves semantics, the benchmarks measure what matters, and LLM-based quality checks approximate human judgment. No new physical entities are postulated; the paper introduces a dataset and a pipeline, not a new mechanism. The only hand-tuned quantities that affect the headline results are the generation temperature and the topic-list size, both ablated.

free parameters (2)

generation temperature = 1.0
Main experiments use temperature 1.0. Ablation in Section 6 shows T=1.0 outperforms T=0.0, 0.5 and 1.2 by up to 3 BLEU; temperatures above 1.2 produce garbage. This is a hyperparameter chosen with knowledge of downstream eval results.
Wikipedia topic list size = 67,573
Data is generated using the full topic list; ablation with a curated 509-topic list reduces fine-tuned BLEU (e.g., LLaMA-3-8B Sundanese from 18.15 to 16.97), showing performance depends on this resource choice.

assumptions (4)

domain assumption The generator LLM can write fluent, topically coherent text in all ten target LRLs with acceptable grammar and script.
Central to the pipeline's data generation step (Section 3); acknowledged in Limitations. If false, the generated target side is poor and backtranslation yields useless parallel data.
domain assumption Backtranslation from LRL to English preserves the semantics of the generated sentences.
Used in Section 3 to create parallel pairs; if the backtranslator mistranslates, the English side is misaligned and fine-tuning learns noise. The paper relies on NLLB-200-3.3B quality in the LRL-to-English direction.
domain assumption FLORES-200 devtest, NTREX-128, and TICO-19 are valid benchmarks for measuring translation quality into these LRLs.
Evaluation in Sections 4 and B.4; the conclusions about improvement are drawn from these benchmarks.
domain assumption LLM-based quality assessment (MetricX-24 QE and topicality judges) correlates with human judgments.
Used in B.6 to validate the generated data's quality and topical alignment in the absence of human ratings.

how reviews work

0 comments

Cite this review

Pith. "Pith review of TopXGen: Topic-Diverse Parallel Data Generation for Low-Resource Machine Translation." pith.science (2026). https://pith.science/paper/TYP47B4K

@misc{pith2026250808680,
  author       = {Pith},
  title        = {Pith review of: TopXGen: Topic-Diverse Parallel Data Generation for Low-Resource Machine Translation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TYP47B4K}},
  note         = {Machine review of arXiv:2508.08680}
}

read the original abstract

LLMs have been shown to perform well in machine translation (MT) with the use of in-context learning (ICL), rivaling supervised models when translating into high-resource languages (HRLs). However, they lag behind when translating into low-resource language (LRLs). Example selection via similarity search and supervised fine-tuning help. However the improvements they give are limited by the size, quality and diversity of existing parallel datasets. A common technique in low-resource MT is synthetic parallel data creation, the most frequent of which is backtranslation, whereby existing target-side texts are automatically translated into the source language. However, this assumes the existence of good quality and relevant target-side texts, which are not readily available for many LRLs. In this paper, we present \textsc{TopXGen}, an LLM-based approach for the generation of high quality and topic-diverse data in multiple LRLs, which can then be backtranslated to produce useful and diverse parallel texts for ICL and fine-tuning. Our intuition is that while LLMs struggle to translate into LRLs, their ability to translate well into HRLs and their multilinguality enable them to generate good quality, natural-sounding target-side texts, which can be translated well into a high-resource source language. We show that \textsc{TopXGen} boosts LLM translation performance during fine-tuning and in-context learning. Code and outputs are available at https://github.com/ArmelRandy/topxgen.

Figures

Figures reproduced from arXiv: 2508.08680 by the authors.

**Figure 1.** Overview of TOPXGEN. An LLM is used to write a diverse set of paragraphs in an LRL guided by topics, example sentences in the LRL and example paragraphs in HRLs. The generated paragraphs are later cleaned and divided into sentences that are backtranslated into the source language to create a sentencelevel parallel dataset. greatly in terms of available resources and representation in NLP research (Joshi et al., 2… view at source ↗

**Figure 2.** TOPXGEN vs SELF-INSTRUCT & KNNINSTRUCT. and FLORES checkpoint by at least 3 BLEU. 10 11 12 13 14 15 BLE U( ) English Hausa 6 10 11 12 13 14 English Igbo 1 2 3 4 5 6 7 8 9 10 15 10 9 8 M etricX-2 4( ) FLORES SmolSent TopXGen 1 2 3 4 5 6 7 8 9 10 18 15 14 13 12 11 [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** TOPXGEN vs FLORES & SMOLSENT. 6 Ablation Studies Impact of the generator and the number of topics. In this section, we focus on Sundanese and study 2 setups. First, we use gpt-4o-mini-2024-07-18 (OpenAI, 2024) as the generator LLM and analyze the performance of small models (LLaMA-2-7B and LLaMA-3-8B) finetuned on 55k sentences. Second, we reduce the number of topics from 67,573 to a curated list of 509 elements a… view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Impact of the temperature of the Generator. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: BLEU and MetricX results for 10 English→X directions from FLORES 200. We fine-tune NLLB-200-3.3B and Gemma-3-27B-PT. We consider 1 model per direction and report the scores (greedy decoding) every 200 steps [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

7 extracted references · 1 canonical work pages

[2]

Association for Compu- tational Linguistics

at EMNLP 2020, Online. Association for Compu- tational Linguistics. Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2018. Unsupervised Statistical Machine Translation. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 3632–3642, Brussels, Belgium. Association for Com- putational Linguistics. Mikel Artetxe, Seb...

work page 2020
[6]

Association for Computational Linguistics

Tagged Back-translation Revisited: Why Does It Really Work? InProceedings of the 58th Annual Meeting of the Association for Computational Lin- guistics, pages 5990–5997, Online. Association for Computational Linguistics. Yasmin Moslem, Rejwanul Haque, John D. Kelleher, and Andy Way. 2023. Adaptive Machine Translation with Large Language Models. InProceedi...

arXiv 2023
[1901]

Franck Burlot and François Yvon

Curran Associates, Inc. Franck Burlot and François Yvon. 2018. Using Mono- lingual Data in Neural Machine Translation: a Sys- tematic Study. InProceedings of the Third Con- ference on Machine Translation: Research Papers, pages 144–155, Brussels, Belgium. Association for Computational Linguistics. Happy Buzaaba, Alexander Wettig, David Ifeoluwa Adelani, a...

arXiv 2018
[2016]

We now have 4-month-old mice that are non-diabetic that used to be diabetic,

Improving Neural Machine Translation Mod- els with Monolingual Data. InProceedings of the 54th Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers), pages 86–96, Berlin, Germany. Association for Computa- tional Linguistics. Shivalika Singh, Freddie Vargus, Daniel D’souza, Börje F. Karlsson, Abinaya Mahendiran, Wei-Yin ...

arXiv 2024
[2019]

InProceedings of the Fourth Conference on Machine Translation (Volume 1: Research Papers), pages 53–63, Florence, Italy

Tagged Back-Translation. InProceedings of the Fourth Conference on Machine Translation (Volume 1: Research Papers), pages 53–63, Florence, Italy. Association for Computational Linguistics. Isaac Caswell, Elizabeth Nielsen, Jiaming Luo, Colin Cherry, Geza Kovacs, Hadar Shemtov, Partha Talukdar, Dinesh Tewari, Baba Mamadi Diane, Koulako Moussa Doumbouya, Dj...

arXiv 2025
[2020]

InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4623–4637, Online

On the Cross-lingual Transferability of Mono- lingual Representations. InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4623–4637, Online. Association for Computational Linguistics. AI at Meta. 2025. The Llama 4 herd: The be- ginning of a new era of natively multimodal AI innovation. https://ai.meta.com/blog...

arXiv 2025
[2024]

Transactions of the Association for Computational Linguistics, 12:979–995

xcomet: Transparent Machine Translation Evaluation through Fine-grained Error Detection. Transactions of the Association for Computational Linguistics, 12:979–995. Suriya Gunasekar, Yi Zhang, Jyoti Aneja, Caio César Teodoro Mendes, Allie Del Giorno, Sivakanth Gopi, Mojan Javaheripi, Piero Kauffmann, Gustavo de Rosa, Olli Saarikivi, Adil Salim, Shital Shah...

arXiv 2023

Pith tools

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

[1] [2]

Association for Compu- tational Linguistics

at EMNLP 2020, Online. Association for Compu- tational Linguistics. Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2018. Unsupervised Statistical Machine Translation. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 3632–3642, Brussels, Belgium. Association for Com- putational Linguistics. Mikel Artetxe, Seb...

work page 2020

[2] [6]

Association for Computational Linguistics

Tagged Back-translation Revisited: Why Does It Really Work? InProceedings of the 58th Annual Meeting of the Association for Computational Lin- guistics, pages 5990–5997, Online. Association for Computational Linguistics. Yasmin Moslem, Rejwanul Haque, John D. Kelleher, and Andy Way. 2023. Adaptive Machine Translation with Large Language Models. InProceedi...

arXiv 2023

[3] [1901]

Franck Burlot and François Yvon

Curran Associates, Inc. Franck Burlot and François Yvon. 2018. Using Mono- lingual Data in Neural Machine Translation: a Sys- tematic Study. InProceedings of the Third Con- ference on Machine Translation: Research Papers, pages 144–155, Brussels, Belgium. Association for Computational Linguistics. Happy Buzaaba, Alexander Wettig, David Ifeoluwa Adelani, a...

arXiv 2018

[4] [2016]

We now have 4-month-old mice that are non-diabetic that used to be diabetic,

Improving Neural Machine Translation Mod- els with Monolingual Data. InProceedings of the 54th Annual Meeting of the Association for Compu- tational Linguistics (Volume 1: Long Papers), pages 86–96, Berlin, Germany. Association for Computa- tional Linguistics. Shivalika Singh, Freddie Vargus, Daniel D’souza, Börje F. Karlsson, Abinaya Mahendiran, Wei-Yin ...

arXiv 2024

[5] [2019]

InProceedings of the Fourth Conference on Machine Translation (Volume 1: Research Papers), pages 53–63, Florence, Italy

Tagged Back-Translation. InProceedings of the Fourth Conference on Machine Translation (Volume 1: Research Papers), pages 53–63, Florence, Italy. Association for Computational Linguistics. Isaac Caswell, Elizabeth Nielsen, Jiaming Luo, Colin Cherry, Geza Kovacs, Hadar Shemtov, Partha Talukdar, Dinesh Tewari, Baba Mamadi Diane, Koulako Moussa Doumbouya, Dj...

arXiv 2025

[6] [2020]

InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4623–4637, Online

On the Cross-lingual Transferability of Mono- lingual Representations. InProceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4623–4637, Online. Association for Computational Linguistics. AI at Meta. 2025. The Llama 4 herd: The be- ginning of a new era of natively multimodal AI innovation. https://ai.meta.com/blog...

arXiv 2025

[7] [2024]

Transactions of the Association for Computational Linguistics, 12:979–995

xcomet: Transparent Machine Translation Evaluation through Fine-grained Error Detection. Transactions of the Association for Computational Linguistics, 12:979–995. Suriya Gunasekar, Yi Zhang, Jyoti Aneja, Caio César Teodoro Mendes, Allie Del Giorno, Sivakanth Gopi, Mojan Javaheripi, Piero Kauffmann, Gustavo de Rosa, Olli Saarikivi, Adil Salim, Shital Shah...

arXiv 2023