Cross-Linguistic Transfer in Multilingual NLP: The Role of Language Families and Morphology

Ajitesh Bankula; Praney Bankula

REVIEW 4 major objections 6 minor 22 references

Cross-Linguistic Transfer in Multilingual NLP: The Role of Language Families and Morphology

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 This paper claims that typological proximity—shared language family and shared morphology—rather than raw pretraining-data size is the dominant driver of zero-shot cross-lingual transfer in XLM-R, based on fine-tuning experiments over 15…

desk verdict The central claim repeats the paper's own references, and the only new evidence—a transfer table with an impossible Chinese diagonal, a POS/NER task mismatch, and a cross-family pair outsoring within-family pairs—cannot support it. read the letter →

arxiv 2505.13908 v1 pith:BVZKD6IM submitted 2025-05-20 cs.CL

Ajitesh Bankula , Praney Bankula This is my paper · ORCID

classification cs.CL

keywords cross-lingualtransfermultilingualNLPlanguagefamiliesmorphologyzero-shotXLM-Rpart-of-speechtaggingtypologicaldistance

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 sets out to show that when a massively multilingual model transfers knowledge to a language it never saw during fine-tuning, success is governed more by how similar the languages are than by how much pretraining data the model had. The authors fine-tune XLM-R on one source language at a time and evaluate zero-shot on the other 14 languages in a 15-language set, using WikiANN annotations for part-of-speech tagging. They report that intra-family source–target pairs usually exceed 0.70 F1 while cross-family pairs often fall below 0.50, and that fusional and agglutinative languages transfer better than isolating ones such as Chinese. The conclusion is that typological proximity, rather than raw training-data size alone, remains the dominant driver of effective knowledge transfer between languages. If true, this gives NLP practitioners a concrete rule for choosing source languages and a reason to build morphology-aware training and tokenization.

What carries the argument

The central object is XLM-R, a 12-layer multilingual Transformer pretrained on about 100 languages with CommonCrawl data; it is the model being fine-tuned, and its learned subword representations are what the paper claims get reused. The argument runs through a transfer matrix built by fine-tuning on each source language's WikiANN training set and evaluating on each target language's test set with no target training sentences. The paper interprets that matrix with three categorical lenses—language family, morphological type (fusional, agglutinative, isolating), and presence of gender marking—plus a distance axis; together these lenses turn raw F1 numbers into evidence that structural affinity, not data volume, drives transfer.

What would settle it

Recompute the same 15-language zero-shot transfer matrix under a single, unambiguous task definition with the source languages and train/test splits specified; if the Chinese diagonal is not near 0.12, or if intra-family advantages shrink once the protocol is fixed, the central comparison collapses. A sharper experiment would hold training data constant and compare genealogically close but typologically distant pairs against genealogically distant but morphologically close pairs (for example Turkish, Finnish, and Tamil, all agglutinative); the paper's claim predicts the morphologically close pairs transfer better regardless of family.

Watch

Extended reading notes

Core claim

On the authors' own terms, the discovery is that XLM-R's zero-shot transfer succeeds to the extent that source and target languages are typologically close: language-family membership and shared morphological features predict F1 more strongly than corpus volume. The evidence is a 15×15 transfer matrix in which intra-family scores, such as Spanish→French, typically pass 0.70, whereas cross-family pairs such as Arabic→Japanese can drop below 0.50. Fusional and agglutinative languages cluster near 0.60–0.65 in intra-family transfer, isolating languages like Chinese near 0.40, and gendered languages beat non-gendered ones by 0.05–0.10 within a family. The paper reads the matrix as showing that shared inflectional systems, affixation patterns, and gender marking enable parameter reuse, while tonal and isolating structure leaves the model without a bridge, so large pretraining corpora cannot erase structural divergence.

Load-bearing premise

The load-bearing premise is that every number in the reported table is a correctly computed zero-shot F1 score for one consistent task, and the paper itself weakens that premise by giving Chinese a diagonal of 0.12, by calling the task POS tagging in one section and NER in another, and by leaving the source 'Base languages' undefined.

Editorial extensions

If this is right

Choosing a source language from the target's own family should improve zero-shot performance; the paper explicitly suggests using Finnish rather than English when building a model for Estonian.
Morphological similarity can partly compensate for genealogical distance, so Turkish and Hungarian—both agglutinative—should transfer better to each other than their family separation alone would predict.
Isolating and tonal languages such as Chinese are the most difficult zero-shot endpoints, so they are the languages most likely to need family-specific models or morphology-aware tokenization.
Injecting morphological information—through analyzers, character-level modeling, or multi-task morphological tagging—should narrow the transfer gap for morphologically complex low-resource languages.
Zero-shot benchmarks should be stratified by family and morphological profile, since overall averages can hide systematic failure on typologically distant languages.

Reading between the lines

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

The paper does not spell out an automated source-selection rule, but its claim implies one: for any low-resource target, pick the source language with the smallest morphological distance rather than the largest corpus, and test this by swapping sources while holding the target fixed.
The paper's distance plot mixes genealogical and morphological distance; separating these with strict typological features would show whether shared agglutinative structure alone (Turkish, Finnish, Tamil) drives transfer without any family tie, which the outlier pattern hints at.
If the claim generalizes beyond these 15 languages, zero-shot evaluations should be reported as family-by-family submatrices; aggregate scores will conceal which languages are being left behind, and model rankings may flip depending on the family mix of the benchmark.

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

What should you know about arXiv:2505.13908? It's an under-specified empirical paper claiming that typological and family proximity drive zero-shot transfer in XLM-R. The claim itself is not new—their own references [11] and [14] state it—and the only new evidence, a pairwise transfer table, is too inconsistent to support anything.

The paper does some things right. The authors have read the right literature: Lauscher, Pires, Ponti, Üstün are all there and cited appropriately. Organizing languages by family and morphology is a sensible frame, and measuring intra- vs cross-family transfer with a fixed model is reasonable. If the table were clean, it would be a small but usable data point.

It is not clean. Section 2.1 says the task is POS tagging, but WikiANN is an NER dataset and Section 2.5 explicitly says "cross-lingual NER performance." The same-language diagonal for Chinese is 0.12, which is impossible for a model fine-tuned and evaluated on Chinese; every other diagonal is above 0.6. German→Arabic (0.686) outscores German→French (0.65), directly contradicting the paper's intra-family claim. The table has no error bars, even though §2.4 says each experiment was run three times. "Base languages" are never defined. The example pair in §2.1 calls English "Indo-Aryan"—it's Indo-European but not Indo-Aryan. "Finish" is used for Finnish, and the table has 17 languages while §2.5 says 15. Figures 1–5 are placeholder captions with no images, so the effect sizes described in §3 cannot be checked.

That list is not a pile-on. Any one of these might be a typo. Together, they mean the central comparison rests on evidence that is internally contradictory. The paper's conclusion—typological proximity dominates transfer—is almost certainly true in the literature, but this manuscript does not add a reliable measurement of it.

The citation pattern is conventional and not self-serving. The writing is rough but readable. The problem is that the empirical core is unusable as written.

Who is this for? Someone teaching a course on multilingual NLP might assign it as an example of how not to report transfer experiments. A researcher would not get a new result from it. I would not send it to peer review in this state; the authors should fix the task description, produce the figures, release code and per-seed results, and reconcile the table with the stated protocol. If they do that, a corrected version could be a minor empirical contribution. As it stands, it's a reject.

Referee Report

4 major / 6 minor

Summary. The paper studies zero-shot cross-lingual transfer with XLM-R across 15 languages from several families, using WikiANN as the data source. The procedure is to fine-tune XLM-R on a source language and evaluate on target languages, reporting a pairwise transfer matrix of average F1 values. The authors claim that intra-family transfer outperforms cross-family transfer, that morphological properties such as fusional or agglutinative structure and gender marking influence transfer, and that linguistic distance correlates negatively with transfer success. The central conclusion, stated in Section 4.1, is that typological proximity, rather than raw training data size alone, remains the dominant driver of effective knowledge transfer between languages.

Significance. If the empirical results were reliable, the paper would add supporting evidence for a claim that is already established in the multilingual NLP literature, for example by Lauscher et al. (2020) and Pires et al. (2019). The paper does not propose a new method, dataset, or theoretical framework, so its significance rests entirely on the credibility of its experiments. A strength is that the authors include their full transfer matrix and state that each experiment was run three times, which makes the empirical basis transparent. However, the reported evidence is internally inconsistent and partly missing, so the central claim is not currently supported. The paper also has no machine-checked proofs or released code; its contribution is purely empirical.

major comments (4)

[§2.1 and §2.5] Section 2.1 defines the task as Part-of-Speech (POS) tagging and names WikiANN as the dataset, but WikiANN is a named-entity recognition dataset and Section 2.5 explicitly refers to 'cross-lingual NER performance.' POS tagging and NER have different label spaces and different sensitivity to morphology, so the reported F1 table cannot be interpreted without knowing which task was actually run. This inconsistency undermines the central comparison because every conclusion in Section 3 depends on the transfer matrix.
[Data table (unnumbered, after References)] The diagonal entry for Chinese (source Chinese, target Chinese) is 0.12, far below every other diagonal value and below many cross-lingual cells; no explanation is offered. Since same-language evaluation should reflect supervised fine-tuning performance, this value is implausible and suggests a data-processing or labeling error. Additionally, the table has rows for Tamil, Korean, and Japanese but no columns for these languages, so it is not a complete pairwise matrix and the missing entries are not discussed.
[§3 and §4.1] Figures 1–5 appear only as placeholder captions; no actual plots are included. The narrative in Sections 3.1–3.4 reports specific effect sizes, including intra-family scores surpassing 0.70, cross-family drops exceeding 0.15, and a negative distance–transfer correlation, none of which can be verified without the figures. Because these figures are the only quantitative support for the claims in Section 4.1, the central conclusion is not supported by the submitted manuscript.
[Data table (unnumbered, after References)] The table contains direct evidence against the 'intra-family dominance' claim: German→Arabic (0.686) is a cross-family pair that outscores German→French (0.65), a within-family pair; similarly, French→Arabic (0.649) is much higher than Arabic→French (0.55). These asymmetries are not discussed. Section 2.4 states that each experiment was run three times, but no error bars, standard deviations, or seed-level results are reported, so the table cannot establish that family or morphology is the dominant factor.

minor comments (6)

[Data table header] The column header 'Finish' should read 'Finnish'; this is a factual error that should be corrected.
[Data table caption] The caption says the table reports 'average F1 Accuracy and Precision,' but it is unclear whether the numbers are F1 scores, accuracy, precision, or some combination; the metric should be stated precisely.
[§2.1] The text labels English as 'Indo-Aryan'; English is Indo-European, and English does not appear in the reported table, so the example pair 'English → German' is misleading.
[§3.3] The text refers to 'geographical (or typological) distance' but never defines the distance metric used in Figure 5; geography and typology are distinct notions and should not be conflated without explanation.
[Throughout] There are numerous typos and grammatical errors, including 'Base languges,' 'out primary reason,' 'we analyzed on15 languages,' and 'compare it to very other language'; the manuscript would benefit from a careful proofreading pass.
[§3] The figures are referenced out of order in the text (Figure 1, Figure 3, Figure 2, Figure 4, Figure 5), which makes the narrative harder to follow and should be renumbered or reordered.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the study is an empirical transfer-matrix analysis with no fitted parameters, no self-citation chain, and no definitional reduction of the conclusion to its inputs.

full rationale

The paper does not derive its headline conclusion from its own definitions or fitted parameters. It reports zero-shot F1 transfer scores between language pairs and compares them across externally defined language-family and morphological categories. The conclusion that 'typological proximity, rather than raw training data size alone, remains the dominant driver of effective knowledge transfer' is an empirical reading of the reported matrix, not an algebraic consequence of how any variable was defined. No parameter is fit to a subset and then renamed as a prediction; no uniqueness theorem from the authors' prior work is invoked; and no ansatz is smuggled in through self-citation. The cited works on linguistic proximity and transferability are independent external studies, and the paper contains no self-citations at all. The tabulated evidence may be internally inconsistent—the task is described as POS tagging in Section 2.1 while Section 2.5 refers to NER, WikiANN is a named-entity dataset, and the Chinese diagonal value of 0.12 is implausibly low—but those are data-validity or correctness concerns, not circularity. Under the hard rule that circularity must be exhibited as a specific reduction by construction or as a fitted input called a prediction, no such step can be quoted from this manuscript. Therefore the honest finding is no significant circularity.

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

The paper relies on external resources (WikiANN, XLM-R, typological classifications) without releasing code or specifying the distance metric. No free parameters were fitted. The main fragility is the correctness of the pairwise scores and the morphological labels.

assumptions (4)

domain assumption WikiANN provides consistently annotated POS data across all 15 languages.
Section 2.1 assumes cross-language comparability of tags, so any label inconsistencies would distort the pairwise scores.
domain assumption XLM-R is a representative multilingual model whose fine-tuning on one language transfers to another.
Section 2.2 selects XLM-R without comparing to other models, so conclusions are model-specific.
domain assumption The language family and morphological feature labels in Table 1 are correct.
Table 1 labels English as Indo-Aryan and lists morphological features by hand; an error here would change the intra-family grouping.
domain assumption The linguistic distance measure used in Section 3.3 is valid.
No definition or source is given for the x-axis in Figure 5, so the claimed distance correlation cannot be audited.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Cross-Linguistic Transfer in Multilingual NLP: The Role of Language Families and Morphology." pith.science (2026). https://pith.science/paper/BVZKD6IM

@misc{pith2026250513908,
  author       = {Pith},
  title        = {Pith review of: Cross-Linguistic Transfer in Multilingual NLP: The Role of Language Families and Morphology},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BVZKD6IM}},
  note         = {Machine review of arXiv:2505.13908}
}

read the original abstract

Cross-lingual transfer has become a crucial aspect of multilingual NLP, as it allows for models trained on resource-rich languages to be applied to low-resource languages more effectively. Recently massively multilingual pre-trained language models (e.g., mBERT, XLM-R) demonstrate strong zero-shot transfer capabilities[14] [13]. This paper investigates cross-linguistic transfer through the lens of language families and morphology. Investigating how language family proximity and morphological similarity affect performance across NLP tasks. We further discuss our results and how it relates to findings from recent literature. Overall, we compare multilingual model performance and review how linguistic distance metrics correlate with transfer outcomes. We also look into emerging approaches that integrate typological and morphological information into model pre-training to improve transfer to diverse languages[18] [19].

Figures

Figures reproduced from arXiv: 2505.13908 by the authors.

**Figure 3.** (A bar chart visualizing the extent of transfer drop across language families.) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

**Figure 2.** (A scatter plot comparing morphological feature impact on intra [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: (A bar chart highlighting the effect of gender markers on transferability.) [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: (A scatter plot demonstrating the correlation between linguistic distance and cross [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 18 canonical work pages

[11]

From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers

Lauscher, Anne, et al. "From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers." EMNLP 2020, pp. 4483–4499. aclanthology.org/2020.emnlp-main.363

work page 2020
[14]

How Multilingual Is Multilingual BERT?

Pires, Telmo, Eva Schlinger, and Dan Garrette. "How Multilingual Is Multilingual BERT?" arXiv, 2019. arxiv.org/abs/1906.01502

arXiv 2019
[1]

Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges

Arivazhagan, Naveen, et al. "Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges." arXiv, 2019. arxiv.org/abs/1907.05019

arXiv 2019
[2]

On the Cross-Lingual Transferability of Monolingual Representations

Artetxe, Mikel, et al. "On the Cross-Lingual Transferability of Monolingual Representations." Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), 2020, pp. 4623–4637. aclanthology.org/2020.acl- main.421

work page 2020
[3]

Unsupervised Cross-Lingual Representation Learning at Scale

Conneau, Alexis, et al. "Unsupervised Cross-Lingual Representation Learning at Scale." arXiv, 2019. arxiv.org/abs/1911.02116

arXiv 2019
[4]

The CoNLL–SIGMORPHON 2018 Shared Task: Universal Morphological Reinflection

Cotterell, Ryan, et al. "The CoNLL–SIGMORPHON 2018 Shared Task: Universal Morphological Reinflection." Proceedings of the CoNLL SIGMORPHON Shared Task, 2018. aclanthology.org/K18-3001

work page 2018
[5]

Gradual Language Model Adaptation Using Fine-Grained Typology

de Vries, Wietse, et al. "Gradual Language Model Adaptation Using Fine-Grained Typology." EMNLP 2021, pp. 321–331. aclanthology.org/2021.emnlp-main.36

work page 2021
[6]

XTREME: A Massively Multilingual Multi-Task Benchmark for Evaluating Cross-Lingual Generalization

Hu, J. Edward, et al. "XTREME: A Massively Multilingual Multi-Task Benchmark for Evaluating Cross-Lingual Generalization." arXiv, 2020. research.google/pubs/pub49314

work page 2020

Show all 22 references

[7]

Google's Multilingual Neural Machine Translation System: Enabling Zero -Shot Translation

Johnson, Melvin, et al. "Google's Multilingual Neural Machine Translation System: Enabling Zero -Shot Translation." Transactions of the Association for Computational Linguistics, vol. 5, 2017, pp. 339–351. aclanthology.org/Q17-1024

2017
[8]

The State and Fate of Linguistic Diversity and Inclusion in the NLP Community

Joshi, Pratik, et al. "The State and Fate of Linguistic Diversity and Inclusion in the NLP Community." ACL 2020, pp. 6282–6293. aclanthology.org/2020.acl-main.560

2020
[9]

Getting the ##Life out of Living: How Adequate Are Word -Pieces for Modelling Complex Morphology?

Klein, Stav, and Reut Tsarfaty. "Getting the ##Life out of Living: How Adequate Are Word -Pieces for Modelling Complex Morphology?" SIGMORPHON 2020, pp. 204–209. aclanthology.org/2020.sigmorphon-1.24

2020
[10]

75 Languages, 1 Model: Parsing Universal Dependencies Universally

Kondratyuk, Dan, and Milan Straka. "75 Languages, 1 Model: Parsing Universal Dependencies Universally." EMNLP 2019, pp. 2779–2795. aclanthology.org/D19-1279

2019
[12]

Language Neutrality in Cross-Lingual Sentence Embeddings

Libovický, Jindřich, et al. "Language Neutrality in Cross-Lingual Sentence Embeddings." EMNLP 2020, pp. 166–

2020
[13]

URIEL and LANG2VEC: Representing Languages as Typological, Geographical, and Phylogenetic Vectors

Littell, Patrick, et al. "URIEL and LANG2VEC: Representing Languages as Typological, Geographical, and Phylogenetic Vectors." EACL 2017, pp. 8–14. aclanthology.org/E17-2002

2017
[15]

Modeling Language Variation and Universals: A Survey on Typological Linguistics for Natural Language Processing

Ponti, Edoardo M., et al. "Modeling Language Variation and Universals: A Survey on Typological Linguistics for Natural Language Processing." Computational Linguistics, vol. 47, no. 2, 2021, pp. 455–500. aclanthology.org/2021.cl-2.3

2021
[16]

The State of Multilingual AI

Ruder, Sebastian. "The State of Multilingual AI." Ruder.io, 2022. ruder.io/state-of-multilingual-ai

2022
[17]

Zero-Shot Dependency Parsing with Pretrained Multilingual Sentence Representations

Tran, Ke, and Arianna Bisazza. "Zero-Shot Dependency Parsing with Pretrained Multilingual Sentence Representations." EMNLP 2019, pp. 3032–3042. aclanthology.org/D19-1301

2019
[18]

Multilingual Dependency Parsing with Typological Features

Üstün, Ahmet, et al. "Multilingual Dependency Parsing with Typological Features." NAACL 2022, pp. 2321–2334. aclanthology.org/2022.naacl-main.168

2022
[19]

LANGSAMP: Language and Script Embeddings for Improved Cross -Lingual Transfer

Wang, Yiren, et al. "LANGSAMP: Language and Script Embeddings for Improved Cross -Lingual Transfer." OpenReview, 2023. openreview.net/forum?id=LANGSAMP2023

2023
[20]

Beto, Bentz, Becas: The Surprising Cross-Lingual Effectiveness of Multilingual BERT

Wu, Shijie, and Mark Dredze. "Beto, Bentz, Becas: The Surprising Cross-Lingual Effectiveness of Multilingual BERT." EMNLP 2019, pp. 833–844. aclanthology.org/D19-1077

2019
[21]

ByT5: Towards a Token-Free Future with Pre-trained Byte-to-Byte Models

Xue, Linting, et al. "ByT5: Towards a Token-Free Future with Pre-trained Byte-to-Byte Models." arXiv, 2021. arxiv.org/abs/2105.13626. Data (average F1 Accuracy and Precision) __________ Arabic Hebrew Indonesian Swahili Chinese Kazakh Turkish Finish Hungarian Arabic 0.75 0.467 ...

2021 arXiv
[178]

aclanthology.org/2020.emnlp-main.13

2020

Pith tools

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

[1] [11]

From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers

Lauscher, Anne, et al. "From Zero to Hero: On the Limitations of Zero-Shot Language Transfer with Multilingual Transformers." EMNLP 2020, pp. 4483–4499. aclanthology.org/2020.emnlp-main.363

work page 2020

[2] [14]

How Multilingual Is Multilingual BERT?

Pires, Telmo, Eva Schlinger, and Dan Garrette. "How Multilingual Is Multilingual BERT?" arXiv, 2019. arxiv.org/abs/1906.01502

arXiv 2019

[3] [1]

Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges

Arivazhagan, Naveen, et al. "Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges." arXiv, 2019. arxiv.org/abs/1907.05019

arXiv 2019

[4] [2]

On the Cross-Lingual Transferability of Monolingual Representations

Artetxe, Mikel, et al. "On the Cross-Lingual Transferability of Monolingual Representations." Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), 2020, pp. 4623–4637. aclanthology.org/2020.acl- main.421

work page 2020

[5] [3]

Unsupervised Cross-Lingual Representation Learning at Scale

Conneau, Alexis, et al. "Unsupervised Cross-Lingual Representation Learning at Scale." arXiv, 2019. arxiv.org/abs/1911.02116

arXiv 2019

[6] [4]

The CoNLL–SIGMORPHON 2018 Shared Task: Universal Morphological Reinflection

Cotterell, Ryan, et al. "The CoNLL–SIGMORPHON 2018 Shared Task: Universal Morphological Reinflection." Proceedings of the CoNLL SIGMORPHON Shared Task, 2018. aclanthology.org/K18-3001

work page 2018

[7] [5]

Gradual Language Model Adaptation Using Fine-Grained Typology

de Vries, Wietse, et al. "Gradual Language Model Adaptation Using Fine-Grained Typology." EMNLP 2021, pp. 321–331. aclanthology.org/2021.emnlp-main.36

work page 2021

[8] [6]

XTREME: A Massively Multilingual Multi-Task Benchmark for Evaluating Cross-Lingual Generalization

Hu, J. Edward, et al. "XTREME: A Massively Multilingual Multi-Task Benchmark for Evaluating Cross-Lingual Generalization." arXiv, 2020. research.google/pubs/pub49314

work page 2020