Neural machine translation often adopts the fine-tuning approach to adapt to specific domains. However, nonrestricted fine-tuning can easily degrade on the general domain and over-fit to the target domain. To mitigate the issue, we propose Prune-Tune, a novel domain adaptation method via gradual pruning. It learns tiny domain-specific sub-networks during fine-tuning on new domains. Prune-Tune alleviates the over-fitting and the degradation problem without model modification. Furthermore, Prune-Tune is able to sequentially learn a single network with multiple disjoint domain-specific sub-networks for multiple domains. Empirical experiment results show that Prune-Tune outperforms several strong competitors in the target domain test set without sacrificing the quality on the general domain in both single and multi-domain settings. The source code and data are available at https://github.com/ohlionel/Prune-Tune.
翻译:神经机器翻译往往采用微调方法适应特定领域。然而,非限制性微调很容易在一般领域降解,并且过于适合目标领域。为了缓解这一问题,我们提议通过逐步修剪来采用新型领域适应方法Prune-Tune。在对新领域进行微调时,它学习了微小的域别子网络。Prune-Tune在不作模型修改的情况下减轻了过度装配和退化问题。此外,Prune-Tune能够按部就班地学习一个单一网络,为多个领域建立多个互不连域专用子网络。经验实验结果显示,Prune-Tune在目标领域测试中优于几个强大的竞争者,而没有牺牲单一和多域环境中的一般领域的质量。源代码和数据见https://github.com/ohlionel/Prune-Tune。