Efficient Bidirectional Neural Machine Translation

08/25/2019
by   Xu Tan, et al.
0

The encoder-decoder based neural machine translation usually generates a target sequence token by token from left to right. Due to error propagation, the tokens in the right side of the generated sequence are usually of poorer quality than those in the left side. In this paper, we propose an efficient method to generate a sequence in both left-to-right and right-to-left manners using a single encoder and decoder, combining the advantages of both generation directions. Experiments on three translation tasks show that our method achieves significant improvements over conventional unidirectional approach. Compared with ensemble methods that train and combine two models with different generation directions, our method saves 50 training time, and also improve inference speed.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
06/23/2019

Sequence Generation: From Both Sides to the Middle

The encoder-decoder framework has achieved promising process for many se...
research
08/16/2019

Attending to Future Tokens For Bidirectional Sequence Generation

Neural sequence generation is typically performed token-by-token and lef...
research
11/05/2019

Improving Bidirectional Decoding with Dynamic Target Semantics in Neural Machine Translation

Generally, Neural Machine Translation models generate target words in a ...
research
09/01/2018

Beyond Error Propagation in Neural Machine Translation: Characteristics of Language Also Matter

Neural machine translation usually adopts autoregressive models and suff...
research
11/01/2018

Language-Independent Representor for Neural Machine Translation

Current Neural Machine Translation (NMT) employs a language-specific enc...
research
02/19/2018

Efficient Implementation of Evaluation Strategies via Token-Guided Graph Rewriting

In implementing evaluation strategies of the lambda-calculus, both corre...
research
02/11/2023

TPE-Net: Track Point Extraction and Association Network for Rail Path Proposal Generation

One essential feature of an autonomous train is minimizing collision ris...

Please sign up or login with your details

Forgot password? Click here to reset