Experts, Errors, and Context: A Large-Scale Study of Human Evaluation for Machine Translation
Open accessThe DOI matches its Crossref record; 38 of the 43 references it cites were found in a registry or an index.
Abstract
Human evaluation of modern high-quality machine translation systems is a difficult problem, and there is increasing evidence that inadequate evaluation procedures can lead to erroneous conclusions. While there has been considerable research on human evaluation, the field still lacks a commonly accepted standard procedure. As a step toward this goal, we propose an evaluation methodology grounded in explicit error analysis, based on the Multidimensional Quality Metrics (MQM) framework. We carry out the largest MQM research study to date, scoring the outputs of top systems from the WMT 2020 shared task in two language pairs using annotations provided by professional translators with access to full document context. We analyze the resulting data extensively, finding among other results a substantially different ranking of evaluated systems from the one established by the WMT crowd workers, exhibiting a clear preference for human over machine output. Surprisingly, we also find that automatic metrics based on pre-trained embeddings can outperform human crowd workers. We make our corpus publicly available for further research.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“These metrics offer notable advantages in that they are (theoretically) fully reproducible, far easier and cheaper to compute, and can avoid some of the issues faced by human studies (Wei and Jia, 2021; Freitag et al., 2021; Amidei et al., 2020).”
From Lessons from the Trenches on Reproducible Evaluation of Language Models (2024) · cites this paper as Freitag et al., 2021
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“While COMET generally aligns better with human judgement than BLEU [2]”
From Metric Inversion in Generative AI Translation: Divergence Between Automated Evaluation Metrics and Human Expert Judgement in the LRM Era (2026) · cites this paper as [2]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
chrF: character n-gram F-score for automatic MT evaluation
Popović, M. (2015) · Proceedings of the Tenth Workshop on Statistical Machine Translation · doi:10.18653/v1/w15-3049 · Cited by 1,220 · cited 1×
“In general, metrics that are not solely based on surface characteristics do somewhat better, though this pattern is not consistent (for example, chrF (Popović, 2015) has a high correlation of 0.”
BackgroundCOMET: A Neural Framework for MT Evaluation
Rei, R., Stewart, C., Farinha, A. C., & Lavie, A. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.213 · Cited by 617
Findings of the 2016 Conference on Machine Translation
Bojar, O., Chatterjee, R., Federmann, C., Graham, Y., Haddow, B., Huck, M., Jimeno Yepes, A., Koehn, P., Logacheva, V., Monz, C., et al. (2016) · Proceedings of the First Conference on Machine Translation: Volume 2, Shared Task Papers · doi:10.18653/v1/w16-2301 · Cited by 539
Achieving Human Parity on Automatic Chinese to English News Translation
Hassan, H., Aue, A., Chen, C., Chowdhary, V., Clark, J. H., Federmann, C., Huang, X., Junczys-Dowmunt, M., Lewis, W. D., Li, M., et al. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.05567 · Cited by 521
Findings of the 2017 Conference on Machine Translation (WMT17)
Bojar, O., Chatterjee, R., Federmann, C., Graham, Y., Haddow, B., Huang, S., Huck, M., Koehn, P., Liu, Q., Logacheva, V., et al. (2017) · Proceedings of the Second Conference on Machine Translation · doi:10.18653/v1/w17-4717 · Cited by 395
Continuous Measurement Scales in Human Evaluation of Machine Translation
Graham, Y., Baldwin, T., Moffat, A., & Zobel, J. (2013) · Trinity's Access to Research Output (TARA) (Trinity College Dublin) · Cited by 353
Has Machine Translation Achieved Human Parity? A Case for Document-level Evaluation
Läubli, S., Sennrich, R., & Volk, M. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1512 · Cited by 283 · cited 1×
“Recent papers have argued that poor human evaluation practices have led to misleading results, including erroneous claims that MT has achieved human parity (Toral, 2020; Läubli et al., 2018).”
BackgroundManual and automatic evaluation of machine translation between European languages
Koehn, P., & Monz, C. (2006) · Proceedings of the Workshop on Statistical Machine Translation - StatMT '06 · doi:10.3115/1654650.1654666 · Cited by 269 · cited 1×
“The first WMT evaluation campaign (Koehn and Monz, 2006) used adequacy and fluency ratings on a 5-point scale acquired from participants as their main metric.”
MethodLommel, A., Uszkoreit, H., & Burchardt, A. (2014) · Tradumàtica tecnologies de la traducció · doi:10.5565/rev/tradumatica.77 · Cited by 233 · cited 1×
“eu) to address the shortcomings of previous quality evaluation methods (Lommel et al., 2014).”
Background
Research works printed but not linked: 1
- ALPAC, 1966, Language and Machines: Computers in Translation and Linguistics; a Report
Cited by
Items 201–280 of 280
Munková, D., Benková, L., Munk, M., Benko, Ľ., & Hájek, P. (2025) · MethodsX · doi:10.1016/j.mex.2025.103613 · Cited by 1
“It has been widely adopted in recent in WMT metrics shared tasks and quality estimation tasks as a high-quality human evaluation method [9 – 11].”
Does Summary Evaluation Survive Translation to Other Languages?
Braun, S., Vasilyev, O., Iskender, N., & Bohannon, J. (2022) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2022.naacl-main.173 · Cited by 1
Test-time Adaptation for Machine Translation Evaluation by Uncertainty Minimization
Zhan, R., Liu, X., Wong, D. F., Zhang, C., Chao, L. S., & Zhang, M. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.47 · Cited by 2
“Specifically, the segment-level MQM score is derived from the count of explicit errors and ranges from -25 to 0, unlike the continuous [1 , 100] scale adopted by WMT (Freitag et al., 2021a).”
BackgroundMethodInfluentialCLSE: Corpus of Linguistically Significant Entities
Chuklin, A., Zhao, J., & Kale, M. (2022) · Workshop on Generation, Evaluation and Metrics (GEM) · doi:10.18653/v1/2022.gem-1.6 · Cited by 4
MMT’s Submission for the WMT 2023 Quality Estimation Shared Task
Wu, Y., Schlegel, V., Beck, D., & Batista-Navarro, R. (2023) · Conference on Machine Translation (WMT) · doi:10.18653/v1/2023.wmt-1.75 · Cited by 1
“Participating systems are assigned the task of predicting the quality score (MQM or DA) for each source-target sentence pair, and their performance is evaluated using Spearman’s rank correlation coefficient as the primary metric, supplemented by the Kendall and Pearson coefficients as secondary metrics for assessment.”
Larionov, D., Viskov, V., Kokush, G., Panchenko, A., & Eger, S. (2023) · Workshop on Evaluation and Comparison of NLP Systems (Eval4NLP) · doi:10.18653/v1/2023.eval4nlp-1.19 · Cited by 1
“The main judgement metric during the competition is segment-level Kendall-τ correlation be-tween model scores and MQM expert annotations.”
MethodInfluentialLearning from Uneven Training Data: Unlabeled, Single Label, and Multiple LabelsPreprint
Zhang, S., Gong, C., & Choi, E. (2021) · arXiv (Cornell University) · Cited by 1
Sun, A., & Han, W. (2025) · Research Square · doi:10.21203/rs.3.rs-7306404/v2 · Cited by 1
JudgeMoE: Distributional Aggregation for LLM-as-a-Judge
Liu, Y., James, J., Wang, Y., Zhao, K., Xiao, C., & Lin, C. (2026) · arXiv · doi:10.48550/arxiv.2610.07109
“Most cells favor JudgeMoE, but TopicalChat and WMT MQM@1–5 show different failure patterns.”
InfluentialWhat Can Unsupervised Machine Translation Contribute to High-Resource Language Pairs?Preprint
Marchisio, K., Freitag, M., & Grangier, D. (2021) · arXiv (Cornell University)
Self-citationHan, L., Sorokina, I., Erofeev, G., & Gladkoff, S. (2021) · arXiv (Cornell University) · Cited by 2
Multilingual Simultaneous Sentence End and Punctuation Prediction (short paper).
Rei, R., Batista, F., Guerreiro, N. M., & Coheur, L. (2021)
A Survey of Machine Translation Technologies for Machine Translation on Patent Documents
Imamura, K., Echizen‐ya, H., Ehara, T., Goto, I., Sudoh, K., Sonoh, S., Tsunakawa, T., Nakazawa, T., Ninomiya, T., & Wang, X. (2022) · Journal of Natural Language Processing · doi:10.5715/jnlp.29.925 · Cited by 1
Evaluer, diagnostiquer et analyser la traduction automatique neuronale
Yvon, F. (2022) · FORUM Revue internationale d’interprétation et de traduction / International Journal of Interpretation and Translation · doi:10.1075/forum.00023.yvo · Cited by 1
Машинный перевод связных текстов: оценка качества и проблемы
Ломоносова, М. И. М. В., & Лукашевич, Н. В. (2023) · doi:10.37892/978-5-6049527-2-6-6
C-XNLI: Croatian Extension of XNLI Dataset
Obadić, L., Jertec, A., Rajnović, M., & Dropuljić, B. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.142 · Cited by 4
“It has been shown that MT models still suffer from errors like mistranslations, non-translations and hallucinations (Freitag et al., 2021; Raunak et al., 2021), which motivated us to analyze the quality of our dataset.”
Liu, Z., Ye, C., Guo, T., Song, D., Gupta, A. C., & Chen, X. (2023) · Proceedings of The World Conference on Social Sciences · doi:10.33422/6th.worldcss.2023.08.120
“For students with disabilities or those from non-native English-speaking backgrounds, MT can be an invaluable tool for accessibility (Freitag et al., 2021).”
BackgroundBriakou, E., Goyal, N., & Carpuat, M. J. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.690 · Cited by 12
Towards Unsupervised Recognition of Token-level Semantic Differences in Related Documents
Vamvas, J., & Sennrich, R. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.835
“…or similarities on the level of individual words has influenced several research areas, notably inter-pretable semantic textual similarity (Lopez-Gazpio et al., 2017) and the evaluation of generated text (Freitag et al., 2021; Fomicheva et al., 2022; Zerva et al., 2022; Rei et al., 2023).”
BackgroundVisual Prediction Improves Zero-Shot Cross-Modal Machine Translation
Hirasawa, T., Bugliarello, E., Elliott, D., & Komachi, M. (2023) · Conference on Machine Translation (WMT) · doi:10.18653/v1/2023.wmt-1.47 · Cited by 4
DOLMACİ, M., & BALKUL, H. İ. (2023) · Karamanoğlu Mehmetbey Üniversitesi Uluslararası Filoloji ve Çeviribilim Dergisi · doi:10.55036/ufced.1392261
A gépi fordítás és kiértékelésének módszerei a fordítástudományban
Yang, Z. G. (2024) · ELTE Digital Institutional Repository (EDIT) (Eötvös Loránd University) · doi:10.21862/kutamodszertan2/14 · Cited by 1
Bevezetés a fordítás és a tolmácsolás kutatásmódszertanába II. Speciális rész
unknown (2024) · ELTE Digital Institutional Repository (EDIT) (Eötvös Loránd University) · doi:10.21862/kutmodszertan2 · Cited by 5
Li, Q. (2024) · University of North Texas Digital Library (University of North Texas) · doi:10.12794/metadc2332575
Pantana, G., & Torre, I. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-4243-1_4
Liang, N. (2024) · Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) · doi:10.1109/ipec61310.2024.00084
Reiter, E. (2024) · Natural Language Generation · doi:10.1007/978-3-031-68582-8_4
Reiter, E. (2024) · Natural Language Generation · doi:10.1007/978-3-031-68582-8_5
Autogenerated MQM Data for Quality Estimation Based on Sequence Labeling
Ge, W., Lei, T., & Gong, Z. (2024) · Lecture notes in computer science · doi:10.1007/978-981-97-9437-9_21
uFOIL: An Unsupervised Fusion of Image Processing and Language Understanding
Rahman, M. A., Hasan, M. T., Howlader, U. F., Kader, M. A., Islam, M. M., Pham, P. H., & Hassan, M. M. (2025) · IEEE Access · doi:10.1109/access.2025.3542417
“As institutions seek to improve operational efficiency and reduce human errors, the demand for automation in document processing has grown significantly [5].”
Li, Y., & Huang, Y. (2025) · Education Reform and Development · doi:10.26689/erd.v7i4.10326 · Cited by 1
“This article demonstrates that MQM error analysis significantly alters system rankings compared to crowd-sourced evaluations and reveals that mistranslation remains the dominant error type in modern NMT outputs [17] .”
Huang, X. (2025) · Lecture Notes in Education Psychology and Public Media · doi:10.54254/2753-7048/2025.cb23801
Sun, A., & Han, W. (2025) · Research Square · doi:10.21203/rs.3.rs-7306404/v1
Huang, Y., Zang, X., Ji, C., & Chen, Z. (2025) · International Journal of Intelligent Systems · doi:10.1155/int/9971702
“Te process necessitates a refnement stage that mimics the critical reasoning employed by human translators during post-editing [8].”
InfluentialDong, X. (2025) · International Conference on Intelligent Mechanical and Human-Computer Interaction Technology (IHCIT) · doi:10.1109/ihcit66787.2025.11199021
“Kocmi & Federmann employed LLMs for error span detection [10] , while Freitag et al. established human-centric evaluation via MQM metrics [11] .”
Meta4BR: Avaliando a Fidelidade Metafórica em Traduções de Metáforas para o Português por LLMs
Stellet, L., Leite, I., Assis, G., & Paes, A. (2025) · Simpósio Brasileiro de Tecnologia da Informação e da Linguagem Humana (STIL) · doi:10.5753/stil.2025.37845
Cai, G., & Zhu, J. (2025) · Lecture notes in computer science · doi:10.1007/978-981-95-2725-0_13 · Cited by 1
Bringing Context into MT Evaluation: Translator Training Insights from the Classroom
Castilho, S. (2025) · Information · doi:10.3390/info16110965
Evaluating Machine Translation Outputs by Comparing Means
Kulhari, A. S., & Joshi, N. (2025) · Lecture notes in networks and systems · doi:10.1007/978-981-96-8796-1_30
Singh, U. P., Garcia, C. A. J., Aisenberg, G. M., Garcia, J. B., Chilatra, J. A. H., Wang, C., de Jesus, D. R., Whalen, E., Canellas, V. S. S., Vargas, A. R. F., et al. (2026) · npj Health Systems · doi:10.1038/s44401-026-00080-5
“Established frameworks suchas the Multidimensional Quality Metrics (MQM) cover accuracy, fl uency, terminology, and adequacy, with error taxonomies also informing clarity, grammar, and overall quality 25 .”
Hu, K., & Gong, W. (2025) · International Conference on AI-enabled Education · doi:10.1145/3768421.3768482
“Second, grammatical dimensions refer to deviations from the structural rules of a language, including syntactic errors, morphological errors, semantic errors, and orthographic errors [8].”
Nguyen, L., Ricatti, F., Davis, S. E., Hunt, J. W., Walker, J. A., Hajek, J., & Travis, C. E. (2025) · Journal of Intercultural Studies · doi:10.1080/07256868.2025.2592305
“Similar discrepancies between human and machine performance have been reported for other ASR and machine translation systems (Spille et al. 2018, Freitag et al. 2021), and are of considerable concern, given that computational studies often rely on narrow benchmarks and quantitative metrics (not…”
Mikros, G., Sosoni, V., Manousakis, V., & Polychroniou, K. (2026) · Journal of Quantitative Linguistics · doi:10.1080/09296174.2026.2612931 · Cited by 1
Munková, D., Forgáč, F., Munk, M., & Hajek, P. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-16851-1_12
Боденчук-Пастухов, Є. В., & Kobylin, I. (2026) · Системи обробки інформації · doi:10.30748/soi.2026.184.12
АДАПТИВНА РОБАСТНА НЕЧІТКА КЛАСТЕРИЗАЦІЯ ПОТОКІВ ДАНИХ НА ОСНОВІ ПРАВДОПОДІБНОГО ПІДХОДУ
Шафроненко, А., Bodyanskiy, Y. V., Шейко, С. О., & Shafronenko, Y. O. (2026) · Системи обробки інформації · doi:10.30748/soi.2026.184.11
Prompting as thinking: Understanding the decisions made by translators in AI-assisted translation
Zhang, S. (2025) · Translation Cognition & Behavior · doi:10.1075/tcb.00099.zha
Traduction automatique de la parole vers des pictogrammes
Macaire, C. (2025) · theses.fr (ABES) · doi:10.70675/06c7e6b3z41b7z467cza019zc899470d6306
Error analysis of translation outputs of the Siswati-English Linguistic Translation Tool
Matfunjwa, M., & Mlambo, R. (2026) · Cogent Arts and Humanities · doi:10.1080/23311983.2026.2651000
Beibei, S. (2026) · Journal of Psycholinguistic Research · doi:10.1007/s10936-026-10274-5
“Freitag et al. (2021 ) conducted a test on a benchmark study to investigate the accuracy and context adequacy of MT, which was evaluated by expert human judg - ment.”
InfluentialAltakarli, R., & Al-Sabbagh, R. (2025) · University of Sharjah (UoS) Journal of Humanities and Social Sciences · doi:10.36394/jhss/22/4/23
Liu, W., & Pagano, A. S. (2026) · Cadernos de Linguagem e Sociedade · doi:10.26512/les.v26i2.59793
LLM-PEA: Leveraging Large Language Models Against Phishing Email Attacks
Hasan, N., Busireddygari, P., Zhao, H., Ren, Y., Xu, J., & Zhang, S. (2026) · International Conference on Dependable Systems and Networks workshops · doi:10.1109/dsn-w70714.2026.00038 · Cited by 2
“Following quality control procedures for machine translation evaluation [18], we removed non-textual emails and kept 179 samples (170 safe, 9 phishing) per language, resulting in 537 total evaluation instances in four languages.”
Machine translation for language access in government settings
Rodríguez, S. A., & Jiménez-Crespo, M. A. (2026) · Digital Translation · doi:10.1075/dt.25018.rod
TransPRM: Process Reward Models Meet Machine Translation via Adaptive Test-Time Compute Scaling
Jiang, X., & Li, T. F. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3417-2_41
Akbari, A. (2026) · Measurement Interdisciplinary Research and Perspectives · doi:10.1080/15366367.2026.2697437
Luo, Z., & Chai, X. (2026) · Applied Sciences · doi:10.3390/app16157505
Junining, E., & Zainuddin, S. Z. (2026) · Cogent Arts and Humanities · doi:10.1080/23311983.2026.2712059
Liu, J. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-27669-8_6
Czulo, O., Šorak, V., & Foradi, M. (2026) · Zeitschrift für Germanistische Linguistik · doi:10.1515/zgl-2026-3004
Wang, X. (2026) · Social Sciences · doi:10.11648/j.ss.20261504.18
Evaluating the Effect of Fine-Tuning on Usability Annotation of English-Slovak Translation
Chotar, M., Kleštinec, M., & Munková, D. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37249-9_19
Hristov, H., Ivanova, V., Ivanov, I., Yasenova, K., & Dimov, G. (2026) · International Conference on Electronics, Engineering Physics and Earth Science (EEPES) · doi:10.3390/engproc2026154058
Intelligent Translation Evaluation and Correction Method for Cross-Language Semantic Alignment
Wang, R., & Wang, S. (2026) · International Conference on Educational Digitalization and Computer Science in the Guangdong-Hong Kong-Macao Greater Bay Area · doi:10.1145/3828183.3828381
“…this paper focuses on the core of cross-lingual semantic alignment and designs a dynamic cross-lingual alignment joint evaluation-correction algorithm (SA-DAC) with enhanced semantic anchors, achieving integrated optimization of semantic alignment, quality assessment, and error correction [3].”
Zhang, Y., & Zhang, Y. (2026) · International Conference on Educational Digitalization and Computer Science in the Guangdong-Hong Kong-Macao Greater Bay Area · doi:10.1145/3828183.3828495
“The gap between automated metrics and human evaluation further exacerbates this issue, as standard metrics may not reliably signal the true quality of AI outputs [5].”
Casamayor-Segarra, A., Ahuir, V., Molina-Marco, A., & Hurtado, L.-F. (2026) · IEEE Access · doi:10.1109/access.2026.3734589
“Following standard practice in human evaluation of machine translation, which relies on adequacy and fluency [53] and on error-based frameworks such as Multidimensional Quality Metrics (MQM) [54], [55], we rated each English–Spanish pair on three criteria: semantic adequacy, Spanish fluency, and…”
Zhao, H. (2026) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-981-95-7622-7_24
Measuring Multi-skill Learning Progress with Optimal Transport
Nurbakova, D. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37982-5_26
GPT-4o Translation of Pediatric Discharge Instructions Across Languages
Kats, D. J., Ray, M., Moorkens, J., Rai, D., Brewster, R., Khan, A., Rader, B. M., Brownstein, J. S., & Hron, J. D. (2026) · Hospital Pediatrics · doi:10.1542/hpeds.2026-009476
Quality-by-Design Framework for Evaluating AI-Driven Machine Translation in Regulatory Affairs
Filatova, D. V., Kadi, H., & Pereme, F. (2026) · Information · doi:10.3390/info17101022
Nguyen, H. H. (2026) · Information · doi:10.3390/info17101026
Wang, Y., Yu, Q., Hui, L., & 彭广陆 (2026) · Research Square · doi:10.21203/rs.3.rs-10122799/v1
Wu, G., Núñez, P. A., Ramos, I., & Hu, Q. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7629414
When is an LLM-as-a-Judge Good Enough?
Wefers, S., & Breitenbach, T. (2026) · The European Journal on Artificial Intelligence · doi:10.1177/30504554261495282
Mattioda, M. M., & Civico, M. (2026) · Terminology International Journal of Theoretical and Applied Issues in Specialized Communication · doi:10.1075/term.00098.mat
Metadata
- Authors
- Freitag, M., Foster, G., Grangier, D., Ratnakar, V., Tan, Q., & Macherey, W.
- Affiliations
- Google (United States) (US)
- Corresponding author
- Freitag, Markus
- Foster, George
- Grangier, David
- Ratnakar, Viresh
- Tan, Qijun
- Macherey, Wolfgang
- ORCID
- Grangier, David
- Venue
- Transactions of the Association for Computational Linguistics
- Publisher
- MIT Press
- Published
- 2021
- Volume
- 9
- Pages
- 1460-1474
- ISSN
- 2307-387X (ISSN-L)
- OpenAlex
- W3159892921
- Cited by
- 271
- References
- 43
- Citations by year
- 2026: 45 · 2025: 48 · 2024: 36 · 2023: 73 · 2022: 45 · 2021: 24
- FWCI
- 15.59
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Locations
- Transactions of the Association for Computational Linguistics (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Social Sciences › Arts and Humanities › Language and Linguistics › Translation Studies and Practices
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Keywords
- human evaluation of machine translation, Multidimensional Quality Metrics, error analysis, evaluation methodology, professional translators, crowd workers, automatic evaluation metrics, pre-trained embeddings, human-machine comparison
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.