Designing Fairly Fair Classifiers Via Economic Fairness Notions

Hossain, Safwan; Mladenovic, Andjela; Shah, Nisarg

doi:10.1145/3366423.3380228

Cited by 23 publications

(30 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is growing interest in making the relationship between fairness in machine learning and social choice theory [25,59,4,20,27,12,16,17], and welfare economics in particular [56,28,34,36,67]. In line with Hu and Chen [28], who focused on classification and parity penalties, we argue that Pareto-efficiency should be part of fairness assessments.…”

Section: Related Workmentioning

confidence: 82%

Two-sided fairness in rankings via Lorenz dominance

Do¹,

Corbett-Davies²,

Atif³

et al. 2021

Preprint

View full text Add to dashboard Cite

We consider the problem of generating rankings that are fair towards both users and item producers in recommender systems. We address both usual recommendation (e.g., of music or movies) and reciprocal recommendation (e.g., dating). Following concepts of distributive justice in welfare economics, our notion of fairness aims at increasing the utility of the worse-off individuals, which we formalize using the criterion of Lorenz efficiency. It guarantees that rankings are Pareto efficient, and that they maximally redistribute utility from better-off to worse-off, at a given level of overall utility. We propose to generate rankings by maximizing concave welfare functions, and develop an efficient inference procedure based on the Frank-Wolfe algorithm. We prove that unlike existing approaches based on fairness constraints, our approach always produces fair rankings. Our experiments also show that it increases the utility of the worse-off at lower costs in terms of overall utility.35th Conference on Neural Information Processing Systems (NeurIPS 2021).

show abstract

Section: Related Workmentioning

confidence: 82%

Two-sided fairness in rankings via Lorenz dominance

Do¹,

Corbett-Davies²,

Atif³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…EO is more effective at reducing the disparities in group-specific error rates while word-embedding debiasing has better performance. Future work can consider more generalized notions of fairness such as preferences-based frameworks, or extend text-specific fairness to contextualized word embeddings (Hossain et al, 2020;Zhang et al, 2020). Further analysis of the fairness performance tradeoff, especially in multimodal settings, will facilitate equitable decision making in the clinical domain.…”

Section: Discussionmentioning

confidence: 99%

Exploring Text Specific and Blackbox Fairness Algorithms in Multimodal Clinical NLP

Chen

Berlot-Attwell²,

Wang³

et al. 2020

Proceedings of the 3rd Clinical Natural Language Processing Workshop

Self Cite

View full text Add to dashboard Cite

Clinical machine learning is increasingly multimodal, collected in both structured tabular formats and unstructured forms such as free text. We propose a novel task of exploring fairness on a multimodal clinical dataset, adopting equalized odds for the downstream medical prediction tasks. To this end, we investigate a modality-agnostic fairness algorithm-equalized odds post processing-and compare it to a text-specific fairness algorithm: debiased clinical word embeddings. Despite the fact that debiased word embeddings do not explicitly address equalized odds of protected groups, we show that a text-specific approach to fairness may simultaneously achieve a good balance of performance and classical notions of fairness. We hope that our paper inspires future contributions at the critical intersection of clinical NLP and fairness. The full source code is available here: https://github.com/ johntiger1/multimodal_fairness

show abstract

“…In the context of fair classification, recent work by Hashimoto et al [21] studies Rawlsian fairness for empirical risk minimization, and observes that Rawlsian fairness prevents disparity amplification over time, which may be unavoidable if we insist on near-equal group-wise performance as a group-fairness constraint. Recent work has also looked at Rawlsian theory to study the veil of ignorance and inequality measurements [18,22,23,34], contextual bandits [26], fair meta-learning [37], envy-free classification [24].…”

Section: Related Workmentioning

confidence: 99%

Rawlsian Fair Adaptation of Deep Learning Classifiers

Shah

Gupta

Deshpande

et al. 2021

Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society

View full text Add to dashboard Cite

Group-fairness in classification aims for equality of a predictive utility across different sensitive sub-populations, e.g., race or gender. Equality or near-equality constraints in group-fairness often worsen not only the aggregate utility but also the utility for the least advantaged sub-population. In this paper, we apply the principles of Pareto-efficiency and least-difference to the utility being accuracy, as an illustrative example, and arrive at the Rawls classifier that minimizes the error rate on the worst-off sensitive sub-population. Our mathematical characterization shows that the Rawls classifier uniformly applies a threshold to an ideal score of features, in the spirit of fair equality of opportunity. In practice, such a score or a feature representation is often computed by a black-box model that has been useful but unfair. Our second contribution is practical Rawlsian fair adaptation of any given black-box deep learning model, without changing the score or feature representation it computes. Given any score function or feature representation and only its second-order statistics on the sensitive sub-populations, we seek a threshold classifier on the given score or a linear threshold classifier on the given feature representation that achieves the Rawls error rate restricted to this hypothesis class. Our technical contribution is to formulate the above problems using ambiguous chance constraints, and to provide efficient algorithms for Rawlsian fair adaptation, along with provable upper bounds on the Rawls error rate. Our empirical results show significant improvement over state-of-the-art group-fair algorithms, even without retraining for fairness. CCS CONCEPTS• Computing methodologies → Supervised learning by classification; • Theory of computation → Machine learning theory; • Social and professional topics → Race and ethnicity; Gender.

show abstract

Designing Fairly Fair Classifiers Via Economic Fairness Notions

Cited by 23 publications

References 16 publications

Two-sided fairness in rankings via Lorenz dominance

Two-sided fairness in rankings via Lorenz dominance

Exploring Text Specific and Blackbox Fairness Algorithms in Multimodal Clinical NLP

Rawlsian Fair Adaptation of Deep Learning Classifiers

Contact Info

Product

Resources

About