Filters








41,567 Hits in 4.2 sec

Variance-Aware Machine Translation Test Sets [article]

Runzhe Zhan, Xuebo Liu, Derek F. Wong, Lidia S. Chao
2021 arXiv   pre-print
We release 70 small and discriminative test sets for machine translation (MT) evaluation called variance-aware test sets (VAT), covering 35 translation directions from WMT16 to WMT20 competitions.  ...  The test sets and the code for preparing variance-aware MT test sets are freely available at https://github.com/NLP2CT/Variance-Aware-MT-Test-Sets .  ...  Introduction Automated machine translation (MT) evaluation relies on metrics and test sets.  ... 
arXiv:2111.04079v1 fatcat:gf43v57z7vdnhon3h7qk2pf4ly

Length-constrained Neural Machine Translation using Length Prediction and Perturbation into Length-aware Positional Encoding

Yui Oka, Katsuhito Sudoh, Satoshi Nakamura
2021 Journal of Natural Language Processing  
Neural machine translation often suffers from an under-translation problem owing to its limited modeling of the output sequence lengths.  ...  In this study, we propose a novel approach to training a Transformer model using length constraints based on length-aware positional encoding (PE).  ...  extended version of the manuscript published in the Proceedings of the 28th International Conference on Computational Linguistics (Oka et al. 2020) , with a further investigation into perturbation length-aware  ... 
doi:10.5715/jnlp.28.778 fatcat:mj2mt6gnezexdorkya2t6m2bjm

Boosting Neural Machine Translation with Dependency-Scaled Self-Attention Network [article]

Ru Peng and Nankai Lin and Yi Fang and Shengyi Jiang and Tianyong Hao and Boyu Chen and Junbo Zhao
2022 arXiv   pre-print
Syntax knowledge contributes its powerful strength in Neural machine translation (NMT) tasks.  ...  To this end, we propose a parameter-free, dependency-scaled self-attention network (Deps-SAN) for syntax-aware Transformer-based NMT.  ...  Translation scores of different NMT models over different length sentences of the test set.  ... 
arXiv:2111.11707v4 fatcat:a5x7zzusvve7npxsmszgnd62va

Enhancing Machine Translation with Dependency-Aware Self-Attention [article]

Emanuele Bugliarello, Naoaki Okazaki
2020 arXiv   pre-print
Most neural machine translation models only rely on pairs of parallel sentences, assuming syntactic information is automatically learned by an attention mechanism.  ...  In this work, we investigate different approaches to incorporate syntactic knowledge in the Transformer model and also propose a novel, parameter-free, dependency-aware self-attention mechanism that improves  ...  The research results have been achieved by "Research and Development of Deep Learning Technology for Advanced Multilingual Speech Translation," the Commissioned Research of National Institute of Information  ... 
arXiv:1909.03149v3 fatcat:ntaicfypjjgnjo5kojkrujshvq

Confidence-Aware Scheduled Sampling for Neural Machine Translation [article]

Yijin Liu, Fandong Meng, Yufeng Chen, Jinan Xu, Jie Zhou
2021 arXiv   pre-print
Scheduled sampling is an effective method to alleviate the exposure bias problem of neural machine translation.  ...  To address this issue, we propose confidence-aware scheduled sampling.  ...  Bridging the gap between training and inference for neural machine translation.  ... 
arXiv:2107.10427v1 fatcat:sxd5kf4hsjagbkh45vupxw5wum

A Productivity Test of Statistical Machine Translation Post-Editing in a Typical Localisation Context

Mirko Plitt, François Masselot
2010 Prague Bulletin of Mathematical Linguistics  
The translation environment recorded translation and post-editing times for each sentence. The results show a productivity increase for each participant, with significant variance across inviduals.  ...  We evaluated the productivity increase of statistical MT post-editing as compared to traditional translation in a two-day test involving twelve participants translating from English to French, Italian,  ...  Test Results Throughput Variance across translators was high. MT allowed all translators to work faster, though in varying proportions: from 20% to 131% 5 .  ... 
doi:10.2478/v10108-010-0010-x fatcat:wab7ci2r5bd5tgno2towng7fee

Detecting and Understanding Generalization Barriers for Neural Machine Translation [article]

Guanlin Li, Lemao Liu, Conghui Zhu, Tiejun Zhao, Shuming Shi
2020 arXiv   pre-print
However, for realistic task like machine translation, the traditional approach measuring generalization in an average sense provides poor understanding for the fine-grained generalization ability.  ...  Based on the modified one, we propose three simple methods for barrier detection by the search-aware risk estimation through counterfactual generation.  ...  Data settings We conduct experiments on Zh⇒En and En⇒Zh translation tasks using the well-known NIST benchmark. The dev and test datasets of the NIST benchmark are marked by year, e.g.  ... 
arXiv:2004.02181v1 fatcat:ik7mnqrg35crhdhn3piw3itira

Improving Anaphora Resolution in Neural Machine Translation Using Curriculum Learning

Dario Stojanovski, Alexander M. Fraser
2019 Machine Translation Summit  
Modeling anaphora resolution is critical for proper pronoun translation in neural machine translation. Recently it has been addressed by context-aware models with varying success.  ...  have access to exactly the same information at test time.  ...  However, Müller et al. (2018) report moderate improvements of the model on their pronoun test set.  ... 
dblp:conf/mtsummit/StojanovskiF19 fatcat:gvuinnpsubef7d3dofrcoqklsu

Computational techniques for increasing PKI policy comprehension by human analysts

Gabriel A. Weaver, Scott Rea, Sean W. Smith
2010 Proceedings of the 9th Symposium on Identity and Trust on the Internet - IDTRUST '10  
Our Citation-Aware HTML enables machines to process human-readable displays of policies in terms of this reference structure.  ...  Our research accelerates PKI operations by enabling machines to translate between policy page numbers and policy reference structure.  ...  The test results are then presented by controlling the styling of our Citation-Aware HTML for the requested policy passage.  ... 
doi:10.1145/1750389.1750396 dblp:conf/idtrust/WeaverRS10 fatcat:2tb7gwi7pbb7dbfum3xln7outa

Machine Translation of Mathematical Text [article]

Aditya Ohri, Tanya Schmah
2020 arXiv   pre-print
We have implemented a machine translation system, the PolyMath Translator, for LaTeX documents containing mathematical text.  ...  The current implementation translates English LaTeX to French LaTeX, attaining a BLEU score of 53.5 on a held-out test corpus of mathematical sentences.  ...  As noted earlier, our main corpus was randomly split into training (80%), validation (10%), and test (10%) sets.  ... 
arXiv:2010.05229v1 fatcat:eve2k74pt5hetnmqglw3ipd2cy

Measuring and Increasing Context Usage in Context-Aware Machine Translation [article]

Patrick Fernandes, Kayo Yin, Graham Neubig, André F. T. Martins
2021 arXiv   pre-print
Using this metric, we measure how much document-level machine translation systems use particular varieties of context.  ...  Recent work in neural machine translation has demonstrated both the necessity and feasibility of using inter-sentential context -- context from sentences other than those currently being translated.  ...  We use the test sets 2011-2014 as validation sets and the 2015 as test sets.  ... 
arXiv:2105.03482v2 fatcat:dz4jsbizsjdnncn5fbqvbne7i4

Forecasting: Adopting the Methodology of Support Vector Machines to Nursing Research

Huey-Ming Tzeng
2006 Worldviews on Evidence-Based Nursing  
Predicting the influence of working motivation and job satisfaction on nurses' intention to quit with a support vector machine: A new approach to set up an early warning in human resource management.  ...  As a common practice, the original dataset usually has to be divided into two groups: (1) the training dataset for training the SVM and (2) the testing dataset for validating the learning machine trained  ... 
doi:10.1111/j.1741-6787.2006.00062.x pmid:16965314 fatcat:5ywwrg6jp5dnhe7desejbaerki

Difficulties in establishing common ground in multiparty groups using machine translation

Naomi Yamashita, Rieko Inaba, Hideaki Kuzuoka, Toru Ishida
2009 Proceedings of the 27th international conference on Human factors in computing systems - CHI 09  
When people communicate in their native languages using machine translation, they face various problems in constructing common ground.  ...  translation.  ...  Where ANOVA is carried out, the test for homogeneity of variance (Levene test) was also carried out. Unless reported, variances were equal between conditions (p>.05).  ... 
doi:10.1145/1518701.1518807 dblp:conf/chi/YamashitaIKI09 fatcat:vjfw6n4tqzgupltdis3gfamfzi

Uncertainty-Aware Machine Translation Evaluation [article]

Taisiya Glushkova, Chrysoula Zerva, Ricardo Rei, André F. T. Martins
2021 arXiv   pre-print
Several neural-based metrics have been recently proposed to evaluate machine translation quality. However, all of them resort to point estimates, which provide limited information at segment level.  ...  We experiment with varying numbers of references and further discuss the usefulness of uncertainty-aware quality estimation (without references) to flag possibly critical translation mistakes.  ...  , where s is source text, t is machine translated text, and R = {r 1 , . . . , r |R| } is a (possibly empty) set of reference translations.  ... 
arXiv:2109.06352v1 fatcat:xlubxd4ynresrawmou4fi4gwim

Privacy-Aware Best-Balanced Multilingual Communication

Mondheera PITUXCOOSUVARN, Takao NAKAGUCHI, Donghui LIN, Toru ISHIDA
2020 IEICE transactions on information and systems  
In machine translation (MT) mediated human-to-human communication, it is not an easy task to select the languages and translation services to be used as the users have various language backgrounds and  ...  Our previous work introduced the best-balanced machine translation mechanism (BBMT) to automatically select the languages and translation services so as to equalize the language barriers of participants  ...  machine translation.  ... 
doi:10.1587/transinf.2019kbp0008 fatcat:k76bksnbhfahrk3o3lkbk7xk7y
« Previous Showing results 1 — 15 out of 41,567 results