On the Impossibility of Non-trivial Accuracy in Presence of Fairness Constraints

Pinzón, Carlos; Palamidessi, Catuscia; Piantanida, Pablo; Valencia, Frank

doi:10.1609/aaai.v36i7.20770

Cited by 4 publications

(7 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, accurate but biased, and fair but faulty predictions do not yield a mutually beneficial trade-off between accuracy and fairness. Such incompatibility has recently been shown in (Pinzón et al 2022) specifically between non-trivial accuracy and equal opportunity, a group fairness criterion.…”

Section: Introductionmentioning

confidence: 91%

“…Accuracy and fairness are both crucial for trustworthy machine learning (Huang et al 2022b,a;Zhang et al 2021;Su et al 2022;Makhlouf, Zhioua, and Palamidessi 2021), but these two aspects may be incompatible fundamentally from their own unilateral perspectives, that is, enhancing one aspect may sacrifice the other inevitably with unacceptable consequences (Dutta et al 2020;Kim, Chen, and Talwalkar 2020;Pinzón et al 2022). For instance, more accurate predictions on loan applicants' incomes can benefit banks with less lending risks, but the underlying ground truth distribution may tend to prefer applicants with the majority or privileged backgrounds, due to historical practices.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Accurate Fairness: Improving Individual Fairness without Trading Accuracy

2023

AAAI

View full text Add to dashboard Cite

Accuracy and individual fairness are both crucial for trustworthy machine learning, but these two aspects are often incompatible with each other so that enhancing one aspect may sacrifice the other inevitably with side effects of true bias or false fairness. We propose in this paper a new fairness criterion, accurate fairness, to align individual fairness with accuracy. Informally, it requires the treatments of an individual and the individual's similar counterparts to conform to a uniform target, i.e., the ground truth of the individual. We prove that accurate fairness also implies typical group fairness criteria over a union of similar sub-populations. We then present a Siamese fairness in-processing approach to minimize the accuracy and fairness losses of a machine learning model under the accurate fairness constraints. To the best of our knowledge, this is the first time that a Siamese approach is adapted for bias mitigation. We also propose fairness confusion matrix-based metrics, fair-precision, fair-recall, and fair-F1 score, to quantify a trade-off between accuracy and individual fairness. Comparative case studies with popular fairness datasets show that our Siamese fairness approach can achieve on average 1.02%-8.78% higher individual fairness (in terms of fairness through awareness) and 8.38%-13.69% higher accuracy, as well as 10.09%-20.57% higher true fair rate, and 5.43%-10.01% higher fair-F1 score, than the state-of-the-art bias mitigation techniques. This demonstrates that our Siamese fairness approach can indeed improve individual fairness without trading accuracy. Finally, the accurate fairness criterion and Siamese fairness approach are applied to mitigate the possible service discrimination with a real Ctrip dataset, by on average fairly serving 112.33% more customers (specifically, 81.29% more customers in an accurately fair way) than baseline models.

show abstract

Section: Introductionmentioning

confidence: 91%

Section: Introductionmentioning

confidence: 99%

Accurate Fairness: Improving Individual Fairness without Trading Accuracy

2023

AAAI

View full text Add to dashboard Cite

show abstract

“…An alternative proof of Corollary 11 that does not use Theorem 10 can be found in the preliminary version of this paper (Pinzón et al, 2022) that was published in the proceedings of AAAI 2022.…”

Section: Necessary and Sufficient Conditionsmentioning

confidence: 99%

“…For reproducibility, we published a repository (Pinzón, 2022) with Python code for generating the figures and algorithms mentioned in this paper, including Algorithms 1 and 2.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the incompatibility of accuracy and equal opportunity

et al. 2023

Self Cite

View full text Add to dashboard Cite

One of the main concerns about fairness in machine learning (ML) is that, in order to achieve it, one may have to trade off some accuracy. To overcome this issue, Hardt et al. (Adv Neural Inf Process Syst 29, 2016) proposed the notion of equality of opportunity (EO), which is compatible with maximal accuracy when the target label is deterministic with respect to the input features. In the probabilistic case, however, the issue is more complicated: It has been shown that under differential privacy constraints, there are data sources for which EO can only be achieved at the total detriment of accuracy, in the sense that a classifier that satisfies EO cannot be more accurate than a trivial (i.e., constant) classifier. In this paper, we strengthen this result by removing the privacy constraint. Namely, we show that for certain data sources, the most accurate classifier that satisfies EO is a trivial classifier. Furthermore, we study the admissible trade-offs between accuracy and EO loss (opportunity difference) and characterize the conditions on the data source under which EO and non-trivial accuracy are compatible.

show abstract

What-is and How-to for Fairness in Machine Learning: A Survey, Reflection, and Perspective

Tang

Zhang

2023

ACM Comput. Surv.

View full text Add to dashboard Cite

We review and reflect on fairness notions proposed in machine learning literature, and make an attempt to draw connections to arguments in moral and political philosophy, especially theories of justice. We survey dynamic fairness inquiries, and further consider the long-term impact induced by current prediction and decision. We present a flowchart that encompasses implicit assumptions and expected outcomes of different fairness inquiries on the data generating process, the predicted outcome, and the induced impact, respectively. We demonstrate the importance of matching the mission (what kind of fairness to enforce) and the means (which appropriate fairness spectrum to analyze) to fulfill the intended purpose.

show abstract

On the Impossibility of Non-trivial Accuracy in Presence of Fairness Constraints

Cited by 4 publications

References 3 publications

Accurate Fairness: Improving Individual Fairness without Trading Accuracy

Accurate Fairness: Improving Individual Fairness without Trading Accuracy

On the incompatibility of accuracy and equal opportunity

What-is and How-to for Fairness in Machine Learning: A Survey, Reflection, and Perspective

Contact Info

Product

Resources

About