Masked Language Models (MLMs) have shown superior performances in numerous downstream NLP tasks when used as text encoders. Unfortunately, MLMs also demonstrate significantly worrying levels of social biases. We show that the previously proposed evaluation metrics for quantifying the social biases in MLMs are problematic due to following reasons: (1) prediction accuracy of the masked tokens itself tend to be low in some MLMs, which raises questions regarding the reliability of the evaluation metrics that use the (pseudo) likelihood of the predicted tokens, and (2) the correlation between the prediction accuracy of the mask and the performance in downstream NLP tasks is not taken into consideration, and (3) high frequency words in the training data are masked more often, introducing noise due to this selection bias in the test cases. To overcome the above-mentioned disfluencies, we propose All Unmasked Likelihood (AUL), a bias evaluation measure that predicts all tokens in a test case given the MLM embedding of the unmasked input. We find that AUL accurately detects different types of biases in MLMs. We also propose AUL with attention weights (AULA) to evaluate tokens based on their importance in a sentence. However, unlike AUL and AULA, previously proposed bias evaluation measures for MLMs systematically overestimate the measured biases, and are heavily influenced by the unmasked tokens in the context.
翻译:语言蒙面模型(MLM)在用作文字编码器时,在许多下游NLP任务中表现优异。不幸的是,MLM公司也表现出令人十分担忧的社会偏见水平。我们表明,先前提出的用于量化MLM公司社会偏见的评价指标存在问题,原因如下:(1) 蒙面图象本身的预测准确性在某些MLM公司中往往较低,这令人怀疑使用(假)预示物的可能性的评价指标的可靠性,以及(2) 蒙面的预测准确性与下游NLP任务业绩的关联性没有被考虑,(3) 培训数据中高频率的单词往往被掩盖,在测试案例中由于这种选择偏好而引起噪音。 为了克服上述不协调性,我们建议所有无序的隐蔽性符号(AUL),这是一个偏差评价措施,在测试案中预测所有象征物的可靠性,因为MLMMMS公司的预测性投入被测量,我们发现AL公司准确检测到MLMS公司中不同种类的偏差性,我们还提议对AUL公司先前的偏重性评价。