Learning to Translate in Real-time with Neural Machine Translation

Gu, Jiatao; Neubig, Graham; Cho, Kyunghyun; Li, Victor O. K.

doi:10.18653/v1/e17-1099

Cited by 158 publications

(243 citation statements)

References 14 publications

Supporting

Mentioning

243

Contrasting

Order By: Relevance

“…With this refined g, we can make several latency metrics content-aware, including average proportion (Cho and Esipova, 2016), consecutive wait (Gu et al, 2017), average lagging , and differentiable average lagging (Arivazhagan et al, 2019b). We opt for differentiable average lagging (DAL) because of its interpretability and because it sidesteps some problems with average lagging (Cherry and Foster, 2019).…”

Section: Latencymentioning

confidence: 99%

“…Furthermore, retranslation's latency-quality trade-off can be manipulated without retraining the base system. It is not the only solution to have these properties; most policies that are not trained jointly with NMT can make the same claims (Cho and Esipova, 2016;Gu et al, 2017;. We conduct an experiment to demonstrate the value of this flexibility, by comparing our Base system to the upgraded Bidi+Beam.…”

Section: Extendability Of Re-translationmentioning

confidence: 99%

“…A typical application is speech translation, where target words must be appended to existing output with no possibility for revision. The corresponding task, which we refer to as streaming translation, has received considerable recent attention, generating custom approaches designed to maximize quality and minimize latency (Cho and Esipova, 2016;Gu et al, 2017;Dalvi et al, 2018;. For certain related applications such as live captioning, * Equal contributions however, the prohibition against revising output is overly stringent.…”

Section: Introductionmentioning

confidence: 99%

“…(3) We test a 0-revision configuration of re-translation, and show that it is surprisingly competitive, due to the effectiveness of data augmentation with prefix pairs. Cho and Esipova (2016) propose the first streaming techniques for NMT, using heuristic agents based on model scores, while Gu et al (2017) extend their work with agents learned using reinforcement learning. recently broke new ground by integrating their read-write agent directly into NMT training.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Re-Translation Strategies for Long Form, Simultaneous, Spoken Language Translation

Arivazhagan

Cherry

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

There has been great progress in improving streaming machine translation, a simultaneous paradigm where the system appends to a growing hypothesis as more source content becomes available. We study a related problem in which revisions to the hypothesis beyond strictly appending words are permitted. This is suitable for applications such as live captioning an audio feed. In this setting, we compare custom streaming approaches to re-translation, a straightforward strategy where each new source token triggers a distinct translation from scratch. We find retranslation to be as good or better than stateof-the-art streaming systems, even when operating under constraints that allow very few revisions. We attribute much of this success to a previously proposed data-augmentation technique that adds prefix-pairs to the training data, which alongside wait-k inference forms a strong baseline for streaming translation. We also highlight re-translation's ability to wrap arbitrarily powerful MT systems with an experiment showing large improvements from an upgrade to its base model.

show abstract

Section: Latencymentioning

confidence: 99%

Section: Extendability Of Re-translationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Re-Translation Strategies for Long Form, Simultaneous, Spoken Language Translation

Arivazhagan

Cherry

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…A number of methods have been proposed to solve this problem and the introduction of neuralnetwork-based MT systems showed a great improvement in simultaneous translation and accuracy of standard (Gu et al 2017). The NMT was proposed by (Kalchbrenner & Blunsom, 2013), (Stuskever et al, 2014) and (Cho et al, 2014a).…”

Section: Neural Machine Translation Systemmentioning

confidence: 99%

Computer Science & Information Technology

2017

View full text Add to dashboard Cite

). The conferences attracted many local and international delegates, presenting a balanced mixture of intellect from the East and from the West.The goal of this conference series is to bring together researchers and practitioners from academia and industry to focus on understanding computer science and information technology and to establish new collaborations in these areas. Authors are invited to contribute to the conference by submitting articles that illustrate research results, projects, survey work and industrial experiences describing significant advances in all areas of computer science and information technology.The NATP-2017, SOFE-2017 Committees rigorously invited submissions for many months from researchers, scientists, engineers, students and practitioners related to the relevant themes and tracks of the workshop. This effort guaranteed submissions from an unparalleled number of internationally recognized top-level researchers. All the submissions underwent a strenuous peer review process which comprised expert reviewers. These reviewers were selected from a talented pool of Technical Committee members and external reviewers on the basis of their expertise. The papers were then reviewed based on their contributions, technical content, originality and clarity. The entire process, which includes the submission, review and acceptance processes, was done electronically. All these efforts undertaken by the Organizing and Technical Committees led to an exciting, rich and a high quality technical conference program, which featured high-impact presentations for all attendees to enjoy, appreciate and expand their expertise in the latest developments in computer network and communications research.In closing, NATP-2017, SOFE-2017 brought together researchers, scientists, engineers, students and practitioners to exchange and share their experiences, new ideas and research results in all aspects of the main workshop themes and tracks, and to discuss the practical challenges encountered and the solutions adopted. The book is organized as a collection of papers from the NATP-2017, SOFE-2017.We would like to thank the General and Program Chairs, organization staff, the members of the Technical Program Committees and external reviewers for their excellent and tireless work. We sincerely wish that all attendees benefited scientifically from the conference and wish them every success in their research. It is the humble wish of the conference organizers that the professional dialogue among the researchers, scientists, engineers, students and educators continues beyond the event and that the friendships and collaborations forged will linger and prosper for many years to come.

show abstract

KLSI Methods for Human Simultaneous Interpretation and Towards Building a Simultaneous Machine Translation System Reflecting the KLSI Methods

Lin

Ming

2021

Artificial Intelligence in HCI

View full text Add to dashboard Cite

Learning to Translate in Real-time with Neural Machine Translation

Cited by 158 publications

References 14 publications

Re-Translation Strategies for Long Form, Simultaneous, Spoken Language Translation

Re-Translation Strategies for Long Form, Simultaneous, Spoken Language Translation

Computer Science & Information Technology

KLSI Methods for Human Simultaneous Interpretation and Towards Building a Simultaneous Machine Translation System Reflecting the KLSI Methods

Contact Info

Product

Resources

About