Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Overview of the dependency-aware self-attention (DASA) mechanism.

More »

Fig 1 Expand

Table 1.

Overall BLEU on four translation benchmarks.

More »

Table 1 Expand

Table 2.

Overall SacreBLEU on four translation benchmarks.

More »

Table 2 Expand

Table 3.

Translation results of different neural machine translation systems.

More »

Table 3 Expand

Table 4.

Ablation on encoder placement of DASA on IWSLT14 De→En.

More »

Table 4 Expand

Table 5.

Sensitivity of translation performance to the bias strength on IWSLT14 De→En.

More »

Table 5 Expand

Table 6.

Mean and standard deviation of BLEU scores over three runs with different random seeds on IWSLT14 De→En.

All other training and decoding settings are identical to the main experiments.

More »

Table 6 Expand

Fig 2.

Low-resource robustness on IWSLT14 De→En under varying fractions of training data.

More »

Fig 2 Expand

Fig 3.

Inference efficiency across datasets on dual RTX 3090Ti GPUs.

More »

Fig 3 Expand

Fig 4.

BLEU by sentence-length bins on IWSLT14 De→En.

More »

Fig 4 Expand

Fig 5.

Visualization of attention distributions in the baseline transformer (left) and the proposed dependency-aware self-attention (DASA, right).

More »

Fig 5 Expand