A user-study on online adaptation of neural machine translation to human post-edits

Karimova, Sariya; Simianer, Patrick; Riezler, Stefan

doi:10.1007/s10590-018-9224-8

Cited by 23 publications

(15 citation statements)

References 40 publications

(52 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…could be improved through humancomputer interaction; a pre-editing protocol was set up to correct the output [8]. Karimova et al (2018) verified the application of M.T. in the field of patent translation and offline M.T.…”

Section: Literature Reviewmentioning

confidence: 81%

“…in the field of patent translation and offline M.T. for different spoken and written languages [9]. Yu et al (2017) combined information technology and MT technology.…”

Section: Literature Reviewmentioning

confidence: 99%

“…is generated, the path with a high probability of high score is selected as the final extended path during the reorganization. The improved maximum probability translation of the non-contiguous phrase decoding algorithm is as shown in (9). 9represents all the phrase pairs under the generalized definition extracted from a sentence pair.…”

Section: B the Extraction Of Discontinuous Phrasesmentioning

confidence: 99%

See 2 more Smart Citations

Application of Data Mining Methods in Internet of Things Technology for the Translation Systems in Traditional Ethnic Books

Luo

Xiang

2020

IEEE Access

View full text Add to dashboard Cite

In order to translate the ethnic classics, based on the research on the Internet of things, machine learning, and translation technology of ethnic classics, the log-linear model is combined with the national corpus scale and the grammatical structure characteristics, and the phrase statistical machine translation is used to establish a discontinuous phrase extraction model. Then, the translation technology is studied from the three aspects of model definition, training, and decoding. Finally, the algorithm is compared with the traditional phrase extraction algorithm to verify its effectiveness. The results show that the extraction number of discontinuous phrase extraction model is significantly higher than that of traditional phrase extraction model, and the model can extract more phrases, handle larger and more complex text, and score higher in translation fluency. From the evaluation indexes scores of Bilingual Evaluation Understudy (B.L.E.U.) and National Institute of Standards and Technology (N.I.S.T.), it can be found that the B.L.E.U. and N.I.S.T. values of the traditional phrase extraction algorithm are lower than those of the discontinuous phrase extraction model algorithm. The discontinuous phrase extraction algorithm can not only extract the regular continuous phrase, but also obtain the discontinuous text, and the translation effect is better. In conclusion, the combination of Internet of things and machine learning can be used in the translation of ethnic classics to achieve high-quality translation of discontinuous phrases, which is of guiding significance for the study of machine translation.

show abstract

Section: Literature Reviewmentioning

confidence: 81%

“…in the field of patent translation and offline M.T. for different spoken and written languages [9]. Yu et al (2017) combined information technology and MT technology.…”

Section: Literature Reviewmentioning

confidence: 99%

Section: B the Extraction Of Discontinuous Phrasesmentioning

confidence: 99%

See 1 more Smart Citation

Application of Data Mining Methods in Internet of Things Technology for the Translation Systems in Traditional Ethnic Books

Luo

Xiang

2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Online adaptation for (neural) machine translation has been thoroughly explored using BLEU (Turchi et al, 2017), simulated keystroke and mouse action ratio (Barrachina et al, 2009) for effort estimation (Peris and Casacuberta, 2018), word prediction accuracy (Wuebker et al, 2016), and user studies (Denkowski et al, 2014;Karimova et al, 2018) (all inter-alia). In (Simianer et al, 2016) immediate adaptation for hierarchical phrase-based MT is specifically investigated, but they also evaluate their systems using humantargeted BLEU and TER.…”

Section: Related Workmentioning

confidence: 99%

Measuring Immediate Adaptation Performance for Neural Machine Translation

Simianer

Wuebker

DeNero

2019

Proceedings of the 2019 Conference of the North

Self Cite

View full text Add to dashboard Cite

Incremental domain adaptation, in which a system learns from the correct output for each input immediately after making its prediction for that input, can dramatically improve system performance for interactive machine translation. Users of interactive systems are sensitive to the speed of adaptation and how often a system repeats mistakes, despite being corrected. Adaptation is most commonly assessed using corpus-level BLEU-or TERderived metrics that do not explicitly take adaptation speed into account. We find that these metrics often do not capture immediate adaptation effects, such as zero-shot and oneshot learning of domain-specific lexical items. To this end, we propose new metrics that directly evaluate immediate adaptation performance for machine translation. We use these metrics to choose the most suitable adaptation method from a range of different adaptation techniques for neural machine translation systems.

show abstract

“…Regarding the NMT technology, several user studies have been recently conducted, analyzing different MT technologies (PB-SMT, NMT and rule-based MT Koponen et al, 2019;Jia et al, 2019) or protocols (IMT versus regular post-editing, Daems and Macken, 2019). Karimova et al (2018) showed savings in human effort, due to the effect of online learning. We also carried a user study which, in contrast to the aforementioned works, involved professional translators.…”

Section: User Studies On Adaptive Machine Translationmentioning

confidence: 99%

Interactivity, Adaptation and Multimodality in Neural Sequence-to-sequence Learning

Abril¹

View full text Add to dashboard Cite

Chapter 1 frames the scope of this thesis, introducing the pattern recognition field and, more specifically, the MT field. It reviews the different historical approaches devised to tackle this problem. Moreover, it sets the experimental framework followed in this thesis and the main scientific objectives. Chapter 2 describes the mathematical model that represents the core of the thesis: neural networks. It addresses the parameter estimation process, describes different neural architectures and a number of techniques used along the thesis to improve the generalization capability of the model. Chapter 3 introduces the neural machine translation technology, describing the most common architectures and decoding process. Moreover, it reviews different aspects relating the NMT field that nowadays receive the attention of the research community. It also compares NMT in the different translation tasks that will be tackled in the thesis. Chapter 4 introduces the interactive-predictive pattern recognition field, that aims to minimize the effort spent by the user while supervising an automatic system. It proposes the application of this theoretical framework to the neural technology, introducing alternative interaction protocols. After that, these interactivepredictive neural systems are evaluated. Chapter 5 describes the adaptation of NMT systems via online learning techniques. After receiving a corrected sample, the system can be updated to include this new knowledge. Here are described the methods to perform this adaptation and introduces two novel alternatives. In addition, an active learning framework for neural systems is proposed, useful for a situation that requires the translation of large amounts of data. All these scenarios are thoroughly evaluated in a variety of conditions, including a user evaluation involving professional post-editors. Chapter 6 departs from the MT problem to tackle different multimodal sequenceto-sequence tasks. More precisely, it is focused on the generation textual descriptions of videos. These techniques are also applied to the captioning of daily events, captured with an egocentric camera. Finally, the interactive-predictive framework described in Chapter 4 is applied to these multimodal systems. Chapter 7 draws the main conclusions of the thesis, describing the scientific contributions and publications derived from it and traces several lines of future research. These chapters are complemented by two appendices. Appendix A describes NMT-Keras, an open-source library developed to build neural models, that has been used to carry out most of the experiments described in the thesis. In Appendix B we provide the results of a survey carried out in the scope of Chapter 5.

show abstract

A user-study on online adaptation of neural machine translation to human post-edits

Cited by 23 publications

References 40 publications

Application of Data Mining Methods in Internet of Things Technology for the Translation Systems in Traditional Ethnic Books

Application of Data Mining Methods in Internet of Things Technology for the Translation Systems in Traditional Ethnic Books

Measuring Immediate Adaptation Performance for Neural Machine Translation

Interactivity, Adaptation and Multimodality in Neural Sequence-to-sequence Learning

Contact Info

Product

Resources

About