Semi-Autoregressive Neural Machine Translation
This addresses the decoding speed problem for machine translation systems, offering a significant improvement over non-autoregressive methods with minimal quality loss.
The paper tackles the slow decoding of autoregressive neural machine translation by proposing a semi-autoregressive Transformer that generates multiple words in parallel, achieving a 5.58x speedup while maintaining 88% translation quality on WMT'14 English-German.
Existing approaches to neural machine translation are typically autoregressive models. While these models attain state-of-the-art translation quality, they are suffering from low parallelizability and thus slow at decoding long sequences. In this paper, we propose a novel model for fast sequence generation --- the semi-autoregressive Transformer (SAT). The SAT keeps the autoregressive property in global but relieves in local and thus is able to produce multiple successive words in parallel at each time step. Experiments conducted on English-German and Chinese-English translation tasks show that the SAT achieves a good balance between translation quality and decoding speed. On WMT'14 English-German translation, the SAT achieves 5.58$\times$ speedup while maintains 88\% translation quality, significantly better than the previous non-autoregressive methods. When produces two words at each time step, the SAT is almost lossless (only 1\% degeneration in BLEU score).