Auditing Fairness under Unobserved Confounding
2403.14713
0
0
✅
Abstract
The presence of inequity is a fundamental problem in the outcomes of decision-making systems, especially when human lives are at stake. Yet, estimating notions of unfairness or inequity is difficult, particularly if they rely on hard-to-measure concepts such as risk. Such measurements of risk can be accurately obtained when no unobserved confounders have jointly influenced past decisions and outcomes. However, in the real world, this assumption rarely holds. In this paper, we show a surprising result that one can still give meaningful bounds on treatment rates to high-risk individuals, even when entirely eliminating or relaxing the assumption that all relevant risk factors are observed. We use the fact that in many real-world settings (e.g., the release of a new treatment) we have data from prior to any allocation to derive unbiased estimates of risk. This result is of immediate practical interest: we can audit unfair outcomes of existing decision-making systems in a principled manner. For instance, in a real-world study of Paxlovid allocation, our framework provably identifies that observed racial inequity cannot be explained by unobserved confounders of the same strength as important observed covariates.
Create account to get full access
Overview
- Measuring unfairness in decision-making systems is challenging, especially when it involves unobserved factors like risk.
- This paper presents a surprising result: meaningful bounds on treatment rates for high-risk individuals can be obtained even without knowing all relevant risk factors.
- The approach leverages data from before any allocation decisions were made to derive unbiased estimates of risk.
- This allows auditing unfair outcomes of existing decision-making systems in a principled manner.
Plain English Explanation
Decision-making systems, like those used to allocate resources or make important choices, can sometimes produce unfair or inequitable outcomes. This research paper explores how to measure this unfairness, even when there are factors that are hard to observe or quantify, like an individual's level of risk.
The key insight is that in many real-world situations, such as the rollout of a new medical treatment, we have data on people's characteristics and outcomes before the decision-making system was put in place. By using this historical data, the researchers show that we can still get a good idea of how fair or unfair the system's decisions are, even if we don't know everything about the individuals involved.
For example, imagine a system that decides who gets a new COVID-19 treatment. The researchers found that they could identify racial inequities in the allocation of this treatment, without needing to know all the details about each person's health or living situation. This is valuable because it allows us to audit and improve decision-making systems to make them more fair and equitable.
Technical Explanation
The key technical insight is that in many real-world settings, such as the release of a new medical treatment, we have access to data on people's characteristics and outcomes from before the decision-making system was put in place. By leveraging this historical data, the researchers show that it is possible to obtain unbiased estimates of individuals' risk levels, even without knowing all the relevant risk factors.
This allows the researchers to calculate meaningful bounds on the treatment rates for high-risk individuals, which serves as a measure of fairness or unfairness in the decision-making system. Importantly, this can be done without relying on the assumption that all relevant risk factors are observed, which is often violated in real-world settings.
The researchers demonstrate the practical utility of this approach with a case study on the allocation of the COVID-19 treatment Paxlovid. They show that their framework can identify racial inequities in the observed treatment rates that cannot be explained by the available observed covariates, suggesting the presence of unfairness in the decision-making process.
Critical Analysis
The researchers acknowledge several caveats and limitations to their approach. First, they note that their method relies on the assumption that the historical data used to estimate risk is representative of the current population, which may not always be the case. Additionally, their framework assumes that the decision-making process can be modeled as a simple treatment assignment problem, which may oversimplify real-world situations.
Another potential concern is that the researchers' approach, while providing bounds on unfairness, does not directly identify the specific sources or mechanisms of unfairness in the decision-making system. More work may be needed to understand and address the underlying causes of inequity.
It is also worth noting that the researchers' approach, while promising, has not been extensively tested or validated across a wide range of real-world decision-making scenarios. Further research and applications of the method will be needed to fully assess its strengths, limitations, and practical implications.
Conclusion
This research paper presents a novel approach to quantifying unfairness in decision-making systems, even when important risk factors are unobserved. By leveraging historical data, the researchers show that meaningful bounds on treatment rates can be obtained, allowing for the principled auditing of existing systems for unfair outcomes.
The potential impact of this work is significant, as it provides a tool for identifying and addressing inequities in high-stakes decision-making, such as the allocation of critical resources or the provision of essential services. As decision-making systems become increasingly prevalent in our society, tools like this will be crucial for ensuring fairness and promoting equitable outcomes.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
🛸
Robust Design and Evaluation of Predictive Algorithms under Unobserved Confounding
Ashesh Rambachan, Amanda Coston, Edward Kennedy
0
0
Predictive algorithms inform consequential decisions in settings where the outcome is selectively observed given choices made by human decision makers. We propose a unified framework for the robust design and evaluation of predictive algorithms in selectively observed data. We impose general assumptions on how much the outcome may vary on average between unselected and selected units conditional on observed covariates and identified nuisance parameters, formalizing popular empirical strategies for imputing missing data such as proxy outcomes and instrumental variables. We develop debiased machine learning estimators for the bounds on a large class of predictive performance estimands, such as the conditional likelihood of the outcome, a predictive algorithm's mean square error, true/false positive rate, and many others, under these assumptions. In an administrative dataset from a large Australian financial institution, we illustrate how varying assumptions on unobserved confounding leads to meaningful changes in default risk predictions and evaluations of credit scores across sensitive groups.
5/21/2024
➖
Hidden yet quantifiable: A lower bound for confounding strength using randomized trials
Piersilvio De Bartolomeis, Javier Abad, Konstantin Donhauser, Fanny Yang
0
0
In the era of fast-paced precision medicine, observational studies play a major role in properly evaluating new treatments in clinical practice. Yet, unobserved confounding can significantly compromise causal conclusions drawn from non-randomized data. We propose a novel strategy that leverages randomized trials to quantify unobserved confounding. First, we design a statistical test to detect unobserved confounding with strength above a given threshold. Then, we use the test to estimate an asymptotically valid lower bound on the unobserved confounding strength. We evaluate the power and validity of our statistical test on several synthetic and semi-synthetic datasets. Further, we show how our lower bound can correctly identify the absence and presence of unobserved confounding in a real-world setting.
5/2/2024
💬
Fairness and Unfairness in Binary and Multiclass Classification: Quantifying, Calculating, and Bounding
Sivan Sabato, Eran Treister, Elad Yom-Tov
0
0
We propose a new interpretable measure of unfairness, that allows providing a quantitative analysis of classifier fairness, beyond a dichotomous fair/unfair distinction. We show how this measure can be calculated when the classifier's conditional confusion matrices are known. We further propose methods for auditing classifiers for their fairness when the confusion matrices cannot be obtained or even estimated. Our approach lower-bounds the unfairness of a classifier based only on aggregate statistics, which may be provided by the owner of the classifier or collected from freely available data. We use the equalized odds criterion, which we generalize to the multiclass case. We report experiments on data sets representing diverse applications, which demonstrate the effectiveness and the wide range of possible uses of the proposed methodology. An implementation of the procedures proposed in this paper and as the code for running the experiments are provided in https://github.com/sivansabato/unfairness.
4/9/2024
🤯
Conformal Counterfactual Inference under Hidden Confounding
Zonghao Chen, Ruocheng Guo, Jean-Franc{c}ois Ton, Yang Liu
0
0
Personalized decision making requires the knowledge of potential outcomes under different treatments, and confidence intervals about the potential outcomes further enrich this decision-making process and improve its reliability in high-stakes scenarios. Predicting potential outcomes along with its uncertainty in a counterfactual world poses the foundamental challenge in causal inference. Existing methods that construct confidence intervals for counterfactuals either rely on the assumption of strong ignorability, or need access to un-identifiable lower and upper bounds that characterize the difference between observational and interventional distributions. To overcome these limitations, we first propose a novel approach wTCP-DR based on transductive weighted conformal prediction, which provides confidence intervals for counterfactual outcomes with marginal converage guarantees, even under hidden confounding. With less restrictive assumptions, our approach requires access to a fraction of interventional data (from randomized controlled trials) to account for the covariate shift from observational distributoin to interventional distribution. Theoretical results explicitly demonstrate the conditions under which our algorithm is strictly advantageous to the naive method that only uses interventional data. After ensuring valid intervals on counterfactuals, it is straightforward to construct intervals for individual treatment effects (ITEs). We demonstrate our method across synthetic and real-world data, including recommendation systems, to verify the superiority of our methods compared against state-of-the-art baselines in terms of both coverage and efficiency
5/22/2024