Later-stage Minimum Bayes-Risk Decoding for Neural Machine Translation

04/11/2017
by   Raphael Shu, et al.
0

For extended periods of time, sequence generation models rely on beam search algorithm to generate output sequence. However, the correctness of beam search degrades when the a model is over-confident about a suboptimal prediction. In this paper, we propose to perform minimum Bayes-risk (MBR) decoding for some extra steps at a later stage. In order to speed up MBR decoding, we compute the Bayes risks on GPU in batch mode. In our experiments, we found that MBR reranking works with a large beam size. Later-stage MBR decoding is shown to outperform simple MBR reranking in machine translation tasks.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
08/10/2021

Sampling-Based Minimum Bayes Risk Decoding for Neural Machine Translation

In neural machine translation (NMT), we search for the mode of the model...
research
07/06/2017

Single-Queue Decoding for Neural Machine Translation

Neural machine translation models rely on the beam search algorithm for ...
research
09/19/2023

MBR and QE Finetuning: Training-time Distillation of the Best and Most Expensive Decoding Methods

Recent research in decoding methods for Natural Language Generation (NLG...
research
10/05/2020

A Streaming Approach For Efficient Batched Beam Search

We propose an efficient batching strategy for variable-length decoding o...
research
02/10/2022

Identifying Weaknesses in Machine Translation Metrics Through Minimum Bayes Risk Decoding: A Case Study for COMET

Neural metrics have achieved impressive correlation with human judgement...
research
10/06/2020

If beam search is the answer, what was the question?

Quite surprisingly, exact maximum a posteriori (MAP) decoding of neural ...
research
12/08/2022

DC-MBR: Distributional Cooling for Minimum Bayesian Risk Decoding

Minimum Bayesian Risk Decoding (MBR) emerges as a promising decoding alg...

Please sign up or login with your details

Forgot password? Click here to reset