In real-world systems, scaling has been critical for improving the translation quality in autoregressive translation (AT), which however has not been well studied for non-autoregressive translation (NAT). In this work, we bridge the gap by systematically studying the impact of scaling on NAT behaviors. Extensive experiments on six WMT benchmarks over two advanced NAT models show that scaling can alleviate the commonly-cited weaknesses of NAT models, resulting in better translation performance. To reduce the side-effect of scaling on decoding speed, we empirically investigate the impact of NAT encoder and decoder on the translation performance. Experimental results on the large-scale WMT20 En-De show that the asymmetric architecture (e.g. bigger encoder and smaller decoder) can achieve comparable performance with the scaling model, while maintaining the superiority of decoding speed with standard NAT models. To this end, we establish a new benchmark by validating scaled NAT models on the scaled dataset, which can be regarded as a strong baseline for future works. We release code, models and system outputs at https://github.com/DeepLearnXMU/Scaling4NAT.

翻译：在实际系统中，规模化对于提升自回归机器翻译（AT）质量至关重要，但非自回归翻译（NAT）的规模化研究尚不充分。本研究通过系统探究规模化对NAT行为的影响，填补了这一空白。基于两种先进NAT模型在六个WMT基准数据集上的大量实验表明：规模化能够缓解NAT模型普遍存在的弱点，从而显著提升翻译性能。为降低规模化对解码速度的副作用，我们通过实验实证探究了NAT编码器与解码器规模对翻译性能的影响。在大型WMT20英德翻译任务上的实验结果显示：非对称架构（如增大编码器、缩小解码器）能在保持标准NAT模型解码速度优势的同时，取得与全规模模型相当的翻译性能。最终，我们通过在大规模数据集上验证规模化NAT模型，建立了新的基准实验，该基准可作为未来研究的强基线。代码、模型及系统输出已开源至 https://github.com/DeepLearnXMU/Scaling4NAT。