FairDistillation: Mitigating Stereotyping in Language Models
This addresses the risk of stereotyping in language models for users and developers, offering a more efficient mitigation approach.
The paper tackles the problem of harmful stereotypes in large pre-trained language models by proposing FairDistillation, a cross-lingual knowledge distillation method that constructs smaller models while controlling for biases, resulting in fairer models at lower cost without significantly harming downstream performance.
Large pre-trained language models are successfully being used in a variety of tasks, across many languages. With this ever-increasing usage, the risk of harmful side effects also rises, for example by reproducing and reinforcing stereotypes. However, detecting and mitigating these harms is difficult to do in general and becomes computationally expensive when tackling multiple languages or when considering different biases. To address this, we present FairDistillation: a cross-lingual method based on knowledge distillation to construct smaller language models while controlling for specific biases. We found that our distillation method does not negatively affect the downstream performance on most tasks and successfully mitigates stereotyping and representational harms. We demonstrate that FairDistillation can create fairer language models at a considerably lower cost than alternative approaches.