Face recognition systems are widely deployed in safety-critical applications, including law enforcement, yet they exhibit bias across a range of socio-demographic dimensions, such as gender and race. Conventional wisdom dictates that model biases arise from biased training data. As a consequence, previous works on bias mitigation largely focused on pre-processing the training data, adding penalties to prevent bias from effecting the model during training, or post-processing predictions to debias them, yet these approaches have shown limited success on hard problems such as face recognition. In our work, we discover that biases are actually inherent to neural network architectures themselves. Following this reframing, we conduct the first neural architecture search for fairness, jointly with a search for hyperparameters. Our search outputs a suite of models which Pareto-dominate all other high-performance architectures and existing bias mitigation methods in terms of accuracy and fairness, often by large margins, on the two most widely used datasets for face identification, CelebA and VGGFace2. Furthermore, these models generalize to other datasets and sensitive attributes. We release our code, models and raw data files at https://github.com/dooleys/FR-NAS.

翻译：人脸识别系统已广泛应用于执法等安全关键领域，但在性别和种族等社会人口维度上仍存在偏见。传统观点认为模型偏见源于有偏的训练数据。因此，先前关于偏见缓解的工作主要集中在预处理训练数据、在训练过程中添加惩罚项以防止偏见影响模型，或对预测结果进行后处理以消除偏见，但这些方法在诸如人脸识别等难题上成效有限。本研究发现，偏见实际上根植于神经网络架构本身。基于这一重新认识，我们首次开展了面向公平性的神经架构搜索，同时联合进行超参数搜索。该搜索输出了一系列模型，它们在准确性和公平性两方面，均以显著优势在最大的人脸识别数据集CelebA和VGGFace2上帕累托支配所有其他高性能架构及现有偏见缓解方法。此外，这些模型还能泛化至其他数据集和敏感属性。我们在https://github.com/dooleys/FR-NAS 公开代码、模型及原始数据文件。