DeepAI AI Chat
Log In Sign Up

Measuring Fairness with Biased Rulers: A Survey on Quantifying Biases in Pretrained Language Models

by   Pieter Delobelle, et al.

An increasing awareness of biased patterns in natural language processing resources, like BERT, has motivated many metrics to quantify `bias' and `fairness'. But comparing the results of different metrics and the works that evaluate with such metrics remains difficult, if not outright impossible. We survey the existing literature on fairness metrics for pretrained language models and experimentally evaluate compatibility, including both biases in language models as in their downstream tasks. We do this by a mixture of traditional literature survey and correlation analysis, as well as by running empirical evaluations. We find that many metrics are not compatible and highly depend on (i) templates, (ii) attribute and target seeds and (iii) the choice of embeddings. These results indicate that fairness or bias evaluation remains challenging for contextualized language models, if not at least highly subjective. To improve future comparisons and fairness evaluations, we recommend avoiding embedding-based metrics and focusing on fairness evaluations in downstream tasks.


On the Intrinsic and Extrinsic Fairness Evaluation Metrics for Contextualized Language Representations

Multiple metrics have been introduced to measure fairness in various nat...

Quantifying Social Biases Using Templates is Unreliable

Recently, there has been an increase in efforts to understand how large ...

HERB: Measuring Hierarchical Regional Bias in Pre-trained Language Models

Fairness has become a trending topic in natural language processing (NLP...

Perturbation Augmentation for Fairer NLP

Unwanted and often harmful social biases are becoming ever more salient ...

Masking as an Efficient Alternative to Finetuning for Pretrained Language Models

We present an efficient method of utilizing pretrained language models, ...

Your fairness may vary: Group fairness of pretrained language models in toxic text classification

We study the performance-fairness trade-off in more than a dozen fine-tu...

Fairness-guided Few-shot Prompting for Large Language Models

Large language models have demonstrated surprising ability to perform in...