Fair Allocation through Selective Information Acquisition

@article{Cai2020FairAT,
  title={Fair Allocation through Selective Information Acquisition},
  author={William Cai and Johann Gaebler and Nikhil Garg and Sharad Goel},
  journal={Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society},
  year={2020}
}
Public and private institutions must often allocate scarce resources under uncertainty. Banks, for example, extend credit to loan applicants based in part on their estimated likelihood of repaying a loan. But when the quality of information differs across candidates (e.g., if some applicants lack traditional credit histories), common lending strategies can lead to disparities across groups. Here we consider a setting in which decision makers---before allocating resources---can choose to spend… 

Figures from this paper

Beyond Reasonable Doubt: Improving Fairness in Budget-Constrained Decision Making using Confidence Thresholds
TLDR
It is shown that a set of carefully chosen confidence thresholds can not only effectively redistribute an information budget according to each individual's needs, but also serve to address individual and group fairness concerns simultaneously.
Dropping Standardized Testing for Admissions Trades Off Information and Access
TLDR
A theoretical framework with testable implications is developed that formalizes the trade-off between the (potentially positive) informational role of a feature and its (negative) exclusionary nature when members of different social groups have unequal access to this feature.
Learning to be Fair: A Consequentialist Approach to Equitable Decision-Making
TLDR
This work develops and describes a method for efficiently learning optimal policies from data for a large family of expressive utility functions, facilitating a more holistic approach to equitable decision-making.
Test-optional Policies: Overcoming Strategic Behavior and Informational Gaps
TLDR
It is found that colleges can overcome two challenges with optional testing: strategic applicants (when those with low test scores can pretend to not have taken the test), and informational gaps (it has more information on those who submit a test score than those who do not).
Standardized Tests and Affirmative Action: The Role of Bias and Variance
TLDR
It is demonstrated how an admissions committee could measure the trade-off in practice to better decide whether to drop their test scores requirement, and if there are substantial barriers to testing, removing the test improves both academic merit and diversity by increasing the size of the applicant pool.
Fairness-aware Online Price Discrimination with Nonparametric Demand Models
TLDR
This work proposes an optimal dynamic pricing policy in terms of regret, which enforces the strict price fairness constraint, and extends the algorithm to a more general notion of fairness, which includes demand fairness as a special case.
Adaptive Sampling Strategies to Construct Equitable Training Datasets
TLDR
The problem of creating equitable training datasets is formalized, and a statistical framework for addressing this problem is proposed, and an adaptive sampling strategy outperforms several common data collection heuristics, including equal and proportional sampling, demonstrating the value of strategic dataset design for building equitable models.
Fair ranking: a critical review, challenges, and future directions
Ranking, recommendation, and retrieval systems are widely used in online platforms and other societal systems, including e-commerce, media-streaming, admissions, gig platforms, and hiring. In the
From Reality to World. A Critical Perspective on AI Fairness
TLDR
This paper provides a new perspective on the debate on AI fairness and shows that criticism of ML unfairness is “realist”, in other words, grounded in an already instituted reality based on demographic categories produced by institutions.
Improving Screening Processes via Calibrated Subset Selection
TLDR
A distribution-free screening algorithm that finds near-optimal shortlists of candidates that contain a desired number of qualified candidates in expectation and it is shown that a variant of the algorithm that calibrates a given classifier multiple times across specific groups can create shortlists with provable diversity guarantees.
...
1
2
...

References

SHOWING 1-10 OF 37 REFERENCES
Fair Algorithms for Learning in Allocation Problems
TLDR
An efficient learning algorithm that converges to an optimal fair allocation even when the allocator does not know the frequency of candidates in each group is proposed, and the performance of the learning algorithm on the Philadelphia Crime Incidents dataset is investigated.
Fair Allocation Rules
Active Fairness in Algorithmic Decision Making
TLDR
active approaches can substantially outperform randomization-based classifiers previously considered optimal, while avoiding limitations such as intra-group unfairness.
Algorithmic Decision Making and the Cost of Fairness
TLDR
This work reformulate algorithmic fairness as constrained optimization: the objective is to maximize public safety while satisfying formal fairness constraints designed to reduce racial disparities, and also to human decision makers carrying out structured decision rules.
On Fairness in Budget-Constrained Decision Making
TLDR
This work provides a framework for choosing a set of stopping criteria that ensures that a probabilistic classifier achieves a single error parity (e.g. equal opportunity) and calibration in budget-constrained decision making.
The Measure and Mismeasure of Fairness: A Critical Review of Fair Machine Learning
TLDR
It is argued that it is often preferable to treat similarly risky people similarly, based on the most statistically accurate estimates of risk that one can produce, rather than requiring that algorithms satisfy popular mathematical formalizations of fairness.
Inherent Trade-Offs in the Fair Determination of Risk Scores
TLDR
Some of the ways in which key notions of fairness are incompatible with each other are suggested, and hence a framework for thinking about the trade-offs between them is provided.
Equality of Opportunity in Supervised Learning
TLDR
This work proposes a criterion for discrimination against a specified sensitive attribute in supervised learning, where the goal is to predict some target based on available features and shows how to optimally adjust any learned predictor so as to remove discrimination according to this definition.
Handbook of Social Choice and Welfare
Social choice theory is concerned with the evaluation of alternative methods of collective decision-making, as well as with the logical foundations of welfare economics. In turn, welfare economics is
Fair prediction with disparate impact: A study of bias in recidivism prediction instruments
TLDR
It is demonstrated that the criteria cannot all be simultaneously satisfied when recidivism prevalence differs across groups, and how disparate impact can arise when an RPI fails to satisfy the criterion of error rate balance.
...
1
2
3
4
...