Abstract

Training data for NLP tasks often exhibits gender bias in that fewer sentences refer to women than to men. In Neural Machine Translation (NMT) gender bias has been shown to reduce translation quality, particularly when the target language has grammatical gender. The recent WinoMT challenge set allows us to measure this effect directly.

Ideally we would reduce system bias by simply debiasing all data prior to training, but this is itself a challenge. Rather than attempt to create a `balanced’ dataset, we adapt to a small set of trusted, gender-balanced examples. This approach gives strong and consistent improvements in gender debiasing with much less computational cost than training from scratch.

A known pitfall of adapting to new domains is `catastrophic forgetting’, which we address both in adaptation and in inference. During adaptation we show that Elastic Weight Consolidation allows a trade-off between general translation quality and bias reduction. During inference we propose a lattice-rescoring scheme which allows extremely strong bias reduction with no degradation of general translation quality. We show this scheme can be applied to reduce gender bias in the output of `black box` online commercial translation systems.