Recent advances in neural methods have led to substantial improvement in the quality of Neural Machine Translation (NMT) systems. However, these systems frequently produce translations with inaccurate gender (Stanovsky et al., 2019), which can be traced to bias in training data. Saunders and Byrne (2020) tackle this problem with a handcrafted dataset containing balanced gendered profession words. By using this data to fine-tune an existing NMT model, they show that gender bias can be significantly mitigated, albeit at the expense of translation quality due to catastrophic forgetting. They recover some of the lost quality with modified training objectives or additional models at inference. We find, however, that simply supplementing the handcrafted dataset with a random sample from the base model training corpus is enough to significantly reduce the catastrophic forgetting. We also propose a novel domain-adaptation technique that leverages in-domain data created with the counterfactual data generation techniques proposed by Zmigrod et al. (2019) to further improve accuracy on the WinoMT challenge test set without significant loss in translation quality. We show its effectiveness in NMT systems from English into three morphologically rich languages French, Spanish, and Italian. The relevant dataset and code will be available at Github.

翻译：神经方法的最近进展显著提升了神经机器翻译（NMT）系统的质量。然而，这些系统经常产生性别翻译不准确的问题（Stanovsky 等，2019），这可以追溯到训练数据中的偏见。Saunders 和 Byrne（2020）通过手工构建一个包含平衡性别职业词汇的数据集来解决这一问题。通过使用该数据对现有 NMT 模型进行微调，他们表明性别偏见可以被显著缓解，尽管由于灾难性遗忘会牺牲翻译质量。他们通过修改训练目标或在推理时使用额外模型来恢复部分丢失的质量。然而，我们发现，仅向手工构建的数据集补充来自基础模型训练语料库的随机样本，就足以显著减少灾难性遗忘。我们还提出了一种新颖的领域适应技术，利用 Zmigrod 等人（2019）提出的反事实数据生成技术创建的领域内数据，在 WinoMT 挑战测试集上进一步提高准确性，且不会显著损失翻译质量。我们展示了从英语到三种形态丰富语言（法语、西班牙语和意大利语）的 NMT 系统中的有效性。相关数据集和代码将在 Github 上提供。