Convolutional Self-Attention Networks
This work addresses the locality limitations in self-attention networks for machine translation, offering an incremental improvement.
The paper tackled the problem of enhancing self-attention networks by strengthening dependencies among neighboring elements and modeling interactions between multiple attention heads, resulting in improved performance in machine translation tasks, outperforming the Transformer baseline and other models without adding parameters.
Self-attention networks (SANs) have drawn increasing interest due to their high parallelization in computation and flexibility in modeling dependencies. SANs can be further enhanced with multi-head attention by allowing the model to attend to information from different representation subspaces. In this work, we propose novel convolutional self-attention networks, which offer SANs the abilities to 1) strengthen dependencies among neighboring elements, and 2) model the interaction between features extracted by multiple attention heads. Experimental results of machine translation on different language pairs and model settings show that our approach outperforms both the strong Transformer baseline and other existing models on enhancing the locality of SANs. Comparing with prior studies, the proposed model is parameter free in terms of introducing no more parameters.