Dynamic Layer Aggregation for Neural Machine Translation with Routing-by-Agreement
This work addresses a specific bottleneck in neural machine translation for improving translation accuracy, representing an incremental advancement.
The paper tackles the problem of static layer aggregation in neural machine translation by proposing a dynamic aggregation method using routing-by-agreement, which consistently outperforms strong baselines on WMT14 English-German and WMT17 Chinese-English datasets.
With the promising progress of deep neural networks, layer aggregation has been used to fuse information across layers in various fields, such as computer vision and machine translation. However, most of the previous methods combine layers in a static fashion in that their aggregation strategy is independent of specific hidden states. Inspired by recent progress on capsule networks, in this paper we propose to use routing-by-agreement strategies to aggregate layers dynamically. Specifically, the algorithm learns the probability of a part (individual layer representations) assigned to a whole (aggregated representations) in an iterative way and combines parts accordingly. We implement our algorithm on top of the state-of-the-art neural machine translation model TRANSFORMER and conduct experiments on the widely-used WMT14 English-German and WMT17 Chinese-English translation datasets. Experimental results across language pairs show that the proposed approach consistently outperforms the strong baseline model and a representative static aggregation model.