Semantics-aware Attention Improves Neural Machine Translation

Slobodkin Aviv, Choshen Leshem, Abend Omri. Arxiv 2021

[Paper]    
Applications Attention Mechanism Model Architecture Pretraining Methods Transformer

The integration of syntactic structures into Transformer machine translation has shown positive results, but to our knowledge, no work has attempted to do so with semantic structures. In this work we propose two novel parameter-free methods for injecting semantic information into Transformers, both rely on semantics-aware masking of (some of) the attention heads. One such method operates on the encoder, through a Scene-Aware Self-Attention (SASA) head. Another on the decoder, through a Scene-Aware Cross-Attention (SACrA) head. We show a consistent improvement over the vanilla Transformer and syntax-aware models for four language pairs. We further show an additional gain when using both semantic and syntactic structures in some language pairs.

Similar Work