An Efficient Transformer Decoder with Compressed Sub-layers

By Yanyang Li, Ye Lin, Tong Xiao, Jingbo Zhu. 2021

33 citations recorded by the source corpus.

The large attention-based encoder-decoder network (Transformer) has become prevailing recently due to its effectiveness. But the high computation complexity of its decoder raises the inefficiency issue. By examining the mathematic formulation of the decoder, we show that under some mild conditions, the architecture could be simplified by compressing its sub-layers, the basic building block of Transformer, and achieves a higher parallelism. We thereby propose Compressed Attention Network, whose decoder layer consists of only one sub-layer instead of three. Extensive experiments on 14 WMT machine translation tasks show that our model is 1.42x faster with performance on par with a strong baseline. This strong baseline is already 2x faster than the widely used standard baseline without loss in performance.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Model Architectures, Efficiency & Compression

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Tao Ge
2. Furu Wei
3. Xin Sun
4. Houfeng Wang
5. Li Dong
6. Yoav Goldberg
7. Jianfeng Gao
8. Xipeng Qiu

In Computer Vision and Pattern Recognition

1. Furu Wei
2. Tao Ge
3. Xin Sun
4. Houfeng Wang
5. Luc Van Gool
6. Dacheng Tao
7. Jianfeng Gao
8. Li Dong

In Model Architectures

1. Xin Sun
2. Houfeng Wang
3. Furu Wei
4. Tao Ge

In Efficiency & Compression

1. Tao Ge
2. Furu Wei
3. Canwen Xu
4. Julian McAuley

Citations in the corpus

• Instantaneous Grammatical Error Correction with Shallow Aggressive Decoding (2021, methodology): “To reduce the computational cost for decoding, we propose to use a shallow decoder, which has proven to be an effective strategy (Kasai et al., 2020; Li et al., 2021) in neural machine translation (NMT), instead of using the Transformer with balanced encoder-decoder depth as the previous…”