Speculative Beam Search for Simultaneous Translation

@inproceedings{Zheng2019SpeculativeBS,
  title={Speculative Beam Search for Simultaneous Translation},
  author={Renjie Zheng and M. Ma and Baigong Zheng and Liang Huang},
  booktitle={EMNLP/IJCNLP},
  year={2019}
}
Beam search is universally used in (full-sentence) machine translation but its application to simultaneous translation remains highly non-trivial, where output words are committed on the fly. In particular, the recently proposed wait-k policy (Ma et al., 2018) is a simple and effective method that (after an initial wait) commits one output word on receiving each input word, making beam search seemingly inapplicable. To address this challenge, we propose a new speculative beam search algorithm… Expand
Simpler and Faster Learning of Adaptive Policies for Simultaneous Translation
TLDR
This work proposes a simple supervised-learning framework to learn an adaptive policy from oracle READ/WRITE sequences generated from parallel text, and shows that this method can learn flexible policies with better BLEU scores and similar latencies compared to previous work. Expand
Simultaneous Translation Policies: From Fixed to Adaptive
TLDR
Experiments show that this algorithm to achieve adaptive policies via a simple heuristic composition of a set of fixed policies can outperform fixed ones by up to 4 BLEU points for the same latency, and more surprisingly, it even surpasses the BLEu score of full-sentence translation in the greedy mode, but with much lower latency. Expand
Opportunistic Decoding with Timely Correction for Simultaneous Translation
TLDR
An opportunistic decoding technique with timely correction ability, which always (over-)generates a certain mount of extra words at each step to keep the audience on track with the latest information to ensure high translation quality. Expand
Simultaneous Multi-Pivot Neural Machine Translation
TLDR
This work proposes multi-pivot translation and applies it to a simultaneous translation setting involving pivot languages and reveals that in a simultaneous pivot NMT setting, using two pivot languages can lead to an improvement of up to 5.8 BLEU. Expand
Alternate Endings: Improving Prosody for Incremental Neural TTS with Predicted Future Text Input
TLDR
The use of predicted future text from a transformer language model can attenuate this loss in a neural TTS system by providing significant improvements over a zero-word lookahead, but only slight gains over random word lookahead. Expand
Fluent and Low-latency Simultaneous Speech-to-Speech Translation with Self-adaptive Training
TLDR
Self-Adaptive Translation (SAT) is proposed which flexibly adjusts the length of translations to accommodate different source speech rates and generates more fluent target speech with substantially lower latency than the baseline, in both Zh En directions. Expand
Fluent and Low-latency Simultaneous Speech-to-Speech Translation with Self-adaptive Training
TLDR
Self-Adaptive Translation is proposed which flexibly adjusts the length of translations to accommodate different source speech rates and generates more fluent target speech latency than the baseline, in both Zh<->En directions. Expand
RealTranS: End-to-End Simultaneous Speech Translation with Convolutional Weighted-Shrinking Transformer
TLDR
RealTranS, an end-to-end model for SST that gradually downsamples the input speech with interleaved convolution and unidirectional Transformer layers for acoustic modeling, and then maps speech features into text space with a weighted-shrinking operation and a semantic encoder is proposed. Expand
Adversarial Machine Learning in Text Analysis and Generation
TLDR
The paper summarizes main research trends in the field such as GAN algorithms, models, types of attacks, and defense against those attacks and concludes with a call for further research into text analysis and generation. Expand
Re-translation versus Streaming for Simultaneous Translation
TLDR
This work finds re-translation to be as good or better than state-of-the-art streaming systems, even when operating under constraints that allow very few revisions. Expand
...
1
2
...

References

SHOWING 1-10 OF 24 REFERENCES
Simpler and Faster Learning of Adaptive Policies for Simultaneous Translation
TLDR
This work proposes a simple supervised-learning framework to learn an adaptive policy from oracle READ/WRITE sequences generated from parallel text, and shows that this method can learn flexible policies with better BLEU scores and similar latencies compared to previous work. Expand
STACL: Simultaneous Translation with Implicit Anticipation and Controllable Latency using Prefix-to-Prefix Framework
TLDR
A novel prefix-to-prefix framework for si- multaneous translation that implicitly learns to anticipate in a single translation model that achieves low latency and reasonable quality (compared to full-sentence translation) on 4 directions. Expand
Learning to Translate in Real-time with Neural Machine Translation
TLDR
A neural machine translation (NMT) framework for simultaneous translation in which an agent learns to make decisions on when to translate from the interaction with a pre-trained NMT environment is proposed. Expand
Monotonic Infinite Lookback Attention for Simultaneous Machine Translation
TLDR
This work presents the first simultaneous translation system to learn an adaptive schedule jointly with a neural machine translation (NMT) model that attends over all source tokens read thus far, and shows that MILk’s adaptive schedule allows it to arrive at latency-quality trade-offs that are favorable to those of a recently proposed wait-k strategy for many latency values. Expand
Learning to Stop in Structured Prediction for Neural Machine Translation
TLDR
A novel ranking method is proposed which enables an optimal beam search stop- ping criteria and a structured prediction loss function is introduced which penalizes suboptimal finished candidates produced by beam search during training. Expand
Don't Until the Final Verb Wait: Reinforcement Learning for Simultaneous Machine Translation
TLDR
A reinforcement learningbased approach to simultaneous machine translation—producing a translation while receiving input words— between languages with drastically different word orders, which shows that the new translation model outperforms batch and monotone translation strategies. Expand
When to Finish? Optimal Beam Search for Neural Text Generation (modulo beam size)
TLDR
A provably optimal beam search algorithm that will always return the optimal-score complete hypothesis (modulo beam size), and finish as soon as the optimality is established is proposed. Expand
Syntax-based Simultaneous Translation through Prediction of Unseen Syntactic Constituents
TLDR
The first experiments applying syntax-based SMT to simultaneous translation are performed, and two methods to prevent degradations in accuracy are proposed: a method to predict unseen syntactic constituents that help generate complete parse trees and a method that waits for more input when the current utterance is not enough to generate a fluent translation. Expand
Incremental Decoding and Training Methods for Simultaneous Translation in Neural Machine Translation
We address the problem of simultaneous translation by modifying the Neural MT decoder to operate with dynamically built encoder and attention. We propose a tunable agent which decides the bestExpand
Correcting Length Bias in Neural Machine Translation
TLDR
It is shown that correcting the brevity problem almost eliminates the beam problem; some commonly-used methods for doing this are compared, finding that a simple per-word reward works well and a simple and quick way to tune this reward using the perceptron algorithm is introduced. Expand
...
1
2
3
...