Towards Equal Opportunity Fairness through Adversarial Learning

@article{Han2022TowardsEO,
  title={Towards Equal Opportunity Fairness through Adversarial Learning},
  author={Xudong Han and Timothy Baldwin and Trevor Cohn},
  journal={ArXiv},
  year={2022},
  volume={abs/2203.06317}
}
Adversarial training is a common approach for bias mitigation in natural language processing. Although most work on debiasing is motivated by equal opportunity, it is not explicitly cap-tured in standard adversarial training. In this paper, we propose an augmented discriminator for adversarial training, which takes the target class as input to create richer features and more explicitly model equal opportunity. Experimental results over two datasets show that our method substantially improves… 

Figures and Tables from this paper

fairlib: A Unified Framework for Assessing and Improving Classification Fairness
This paper presents fairlib , an open-source framework for assessing and improving classification fairness. It provides a systematic framework for quickly reproducing existing baseline models,

References

SHOWING 1-10 OF 29 REFERENCES
What’s in a Domain? Learning Domain-Robust Text Representations using Adversarial Training
TLDR
This work proposes a novel method to optimise both in- and out-of-domain accuracy based on joint learning of a structured neural model with domain-specific and domain-general components, coupled with adversarial training for domain.
FairBatch: Batch Selection for Model Fairness
TLDR
The batch selection algorithm, which the authors call FairBatch, implements this optimization and supports prominent fairness measures: equal opportunity, equalized odds, and demographic parity and is compatible with existing batch selection techniques intended for different purposes, thus gracefully achieving multiple purposes.
Null It Out: Guarding Protected Attributes by Iterative Nullspace Projection
TLDR
This work presents Iterative Null-space Projection (INLP), a novel method for removing information from neural representations based on repeated training of linear classifiers that predict a certain property the authors aim to remove, followed by projection of the representations on their null-space.
Balancing out Bias: Achieving Fairness Through Training Reweighting
TLDR
This paper introduces a very simple but highly effective method for countering bias using instance reweighting, based on the frequency of both task labels and author demographics, and extends the method in the form of a gated model which incorporates the author demographic as an input.
Decoupling Adversarial Training for Fair NLP
TLDR
This paper proposes a training strategy which needs only a small volume of protected labels in adversarial training, incorporating an estimation method to transfer private-labelled instances from one dataset to another.
Diverse Adversaries for Mitigating Bias in Training
TLDR
Experimental results show that the proposed novel approach to adversarial learning based on the use of multiple diverse discriminators substantially improves over standard adversarial removal methods, in terms of reducing bias and stability of training.
fairlib: A Unified Framework for Assessing and Improving Classification Fairness
This paper presents fairlib , an open-source framework for assessing and improving classification fairness. It provides a systematic framework for quickly reproducing existing baseline models,
Evaluating Debiasing Techniques for Intersectional Biases
TLDR
It is argued that a truly fair model must consider ‘gerrymandering’ groups which comprise not only single attributes, but also intersectional groups, and an extension of the iterative nullspace projection technique which can handle multiple identities is evaluated.
Fairness without Demographics through Adversarially Reweighted Learning
TLDR
The proposed Adversarially Reweighted Learning (ARL) hypothesizes that non-protected features and task labels are valuable for identifying fairness issues, and can be used to co-train an adversarial reweighting approach for improving fairness.
Conditional Learning of Fair Representations
TLDR
It is demonstrated both in theory and on two real-world experiments that the proposed algorithm leads to a better utility-fairness trade-off on balanced datasets compared with existing algorithms on learning fair representations.
...
...