Fairness Constraints Research Articles

Bias mitigation for fair automation of classification tasks

AbstractThe incorporation of machine learning algorithms into high‐risk decision‐making tasks has raised some alarms in the scientific community. Research shows that machine learning‐based technologies can contain biases that cause unfair decisions for certain population groups. The fundamental danger of ignoring this problem is that machine learning methods can not only reflect the biases present in our society but could also amplify them. This article presents the design and validation of a technology to assist the fair automation of classification problems. In essence, the proposal is based on taking advantage of the intermediate solutions generated during the resolution of classification problems through using Auto‐ML tools, in particular, AutoGOAL, to create unbiased/fair classifiers. The technology employs a multi‐objective optimization search to find the collection of models with the best trade‐offs between performance and fairness. To solve the optimization problem, we introduce a combination of Probabilistic Grammatical Evolution Search and NSGA‐II. The technology was evaluated using the Adult dataset from the UCI repository, a common benchmark in related research. Results were compared with other published results in scenarios with single and multiple fairness definitions. Our experiments demonstrate the technology's ability to automate classification tasks while incorporating fairness constraints. Additionally, our method achieves competitive results against other bias mitigation techniques. A notable advantage of our approach is its minimal requirement for machine learning expertise, thanks to its Auto‐ML foundation. This makes the technology accessible and valuable for advancing fairness in machine learning applications. The source code is available online for the research community.

Perspectives of Defining Algorithmic Fairness in Customer-oriented Applications: A Systematic Literature Review

Automated decision-making systems are massively engaged in different types of businesses, including customer-oriented sectors, and bring countless achievements in persuading customers with more personalized experiences. However, it was observed that the decisions made by the algorithms could bring unfairness to a person or a group of people, according to recent studies. Thus, algorithmic fairness has become a spotlight research area, and defining a concrete version of fairness notions has also become significant research. In existing literature, there are more than 21 definitions of algorithmic fairness. Many studies have shown that each notion has an incompatibility problem, and it is still necessary to make those notions more adaptable to the legal and social principles of the desired sectors. Yet, the constraints of algorithmic fairness for customer-oriented areas have not been thoroughly studied. This motivates us to work on a systematic literature review to investigate the sectors concerned about algorithmic fairness as a significant matter when using machine-based decision-making systems, what are the well-applied algorithmic fairness notions, and why they can or cannot be directly applicable to the customer-oriented sectors, what are the possible algorithmic fairness constraints for the customer-oriented sectors. By applying the standard guidelines of systematic literature review, we explored 65 prominent articles thoroughly. The findings show 43 different ways of algorithmic fairness notions in the varieties of domains. We also identified the three important perspectives to be considered for enhancing algorithmic fairness notions in the customer-oriented sectors.

Approximation algorithm for prize-collecting vertex cover with fairness constraints

Approximation algorithm for prize-collecting vertex cover with fairness constraints

A Benchmark Proposal for Non‐Generative Fair Adversarial Learning Strategies Using a Fairness‐Utility Trade‐off Metric

ABSTRACTAI systems for decision‐making have become increasingly popular in several areas. However, it is possible to identify biased decisions in many applications, which have become a concern for the computer science, artificial intelligence, and law communities. Therefore, researchers are proposing solutions to mitigate bias and discrimination among decision‐makers. Some explored strategies are based on GANs to generate fair data. Others are based on adversarial learning to achieve fairness by encoding fairness constraints through an adversarial model. Moreover, it is usual for each proposal to assess its model with a specific metric, making comparing current approaches a complex task. Therefore, this work proposes a systematical benchmark procedure to assess the fair machine learning models. The proposed procedure comprises a fairness‐utility trade‐off metric (), the utility and fairness metrics to compose this assessment, the used datasets and preparation, and the statistical test. A previous work presents some of these definitions. The present work enriches the procedure by increasing the applied datasets and statistical guarantees when comparing the models' results. We performed this benchmark evaluation for the non‐generative adversarial models, analyzing the literature models from the same metric perspective. This assessment could not indicate a single model which better performs for all datasets. However, we built an understanding of how each model performs on each dataset with statistical confidence.

Joint Trajectory and Power Optimization for Energy-Efficient UAV Redeployment Against an Eavesdropper Under Consistent Fairness Constraint

Joint Trajectory and Power Optimization for Energy-Efficient UAV Redeployment Against an Eavesdropper Under Consistent Fairness Constraint

Spectral clustering with scale fairness constraints

Spectral clustering with scale fairness constraints

Fair Incentives for Repeated Engagement

We study a decision-maker’s problem of finding optimal monetary incentive schemes for retention when faced with agents whose participation decisions (stochastically) depend on the incentive they receive. Our focus is on policies constrained to fulfill two fairness properties that preclude outcomes wherein different groups of agents experience different treatment on average. We formulate the problem as a high-dimensional stochastic optimization problem and study it through the use of a closely related deterministic variant. We show that the optimal static solution to this deterministic variant is asymptotically optimal for the dynamic problem under fairness constraints. Though solving for the optimal static solution gives rise to a nonconvex optimization problem, we uncover a structural property that allows us to design a tractable, fast-converging heuristic policy. Traditional schemes for retention ignore fairness constraints; indeed, the goal in these is to use differentiation to incentivize repeated engagement with the system. Our work (i) shows that even in the absence of explicit discrimination, dynamic policies may unintentionally discriminate between agents of different types by varying the type composition of the system, and (ii) presents an asymptotically optimal policy to avoid such discriminatory outcomes.

FairColor: An efficient algorithm for the Balanced and Fair Reviewer Assignment Problem

As the volume of submitted papers continues to rise, ensuring a fair and accurate assignment of manuscripts to reviewers has become increasingly critical for academic conference organizers. Given the paper-reviewer similarity scores, this study introduces the Balanced and Fair Reviewer Assignment Problem (BFRAP), which aims to maximize the overall similarity score (efficiency) and the minimum paper score (fairness) subject to coverage, load balance, and fairness constraints. Addressing the challenges posed by these constraints, we conduct a theoretical investigation into the threshold conditions for the problem’s feasibility and optimality. To facilitate this investigation, we establish a connection between BFRAP, defined over m reviewers, and the Equitable m-Coloring Problem. Building on this theoretical foundation, we propose FairColor, an algorithm designed to retrieve fair and efficient assignments. We compare FairColor to Fairflow and FairIR, two state-of-the-art algorithms designed to find fair assignments under similar constraints. Empirical experiments were conducted on four real and two synthetic datasets involving (paper, reviewer) matching scores ranging from (100,100) to (10124,5880). Results demonstrate that FairColor is able to find efficient and fair assignments quickly compared to Fairflow and FairIR. Notably, in the largest instance involving 10,124 manuscripts and 5680 reviewers, FairColor retrieves fair and efficient assignments in just 67.64 s. This starkly contrasts both other methods, which require significantly longer computation times (45 min for Fairflow and 3 h 24 min for FairIR), even on more powerful machines. These results underscore FairColor as a promising alternative to current state-of-the-art assignment techniques.

Optimal Design of Raising Retirement Age

Population aging threatens the future fiscal balance of the public pension system across many countries, and policymakers have often undertaken parametric pension reform as a solution. However, any reform will inevitably introduce intergenerational conflicts and thus needs a gradual transition to allow individuals to slowly adapt to their new retirement plans. In this article, we study the optimal schedule for lifting the full retirement age using the optimal control approach. Given the aggregate actuarial loss of the pension system, our target is to smoothly distribute the deficit across different generations, subject to a fairness constraint that prohibits any cohorts from taking advantage of others. We further illustrate our results using China as an empirical example and investigate the impacts of different actuarial assumptions, benefit accrual formulae, and population structures.

CA-Live360: Crowd-assisted transcoding and delivery for live 360-degree video streaming

In the burgeoning field of digital media, the popularity of live 360-degree video has witnessed an upward trajectory, underpinned by its immersive experience. In live 360-degree video streaming services, transcoding enormous video tiles and providing low-latency services for global viewers require significantly more computing and communication resources than traditional live services. Inspired by the huge crowdsourcing computational resources available at crowd, this paper designs CA-Live360, a Crowd-Assisted Live 360-degree video streaming solution, in which transcoding and delivery are supported by crowdsourcing devices. Due to the dynamic availability of resources, it is challenging to achieve low-latency live services, while guaranteeing fair transcoding. Therefore, firstly an innovative fairness-guaranteed transcoding task assignment scheme is proposed. This scheme employs a novel Fair Bandit algorithm based on the combinatorial multi-armed bandit approach to achieve low-latency transcoding while considering time-varying available resources and fairness constraints. Secondly, to reduce delivery latency, we design a transcoding-aware delivery scheme, in which a provider–requester matching algorithm is introduced to realize effective requester scheduling. Real-world trace-driven experiments demonstrate the improved performance of the proposed CA-Live360 solution in terms of system delay and fairness.

A deep learning approach for fairness-based time of use tariff design

Time of use (TOU) tariffs aim to shape demand by reducing peak demand. This is significant given the increasing pressure on peak energy supply, particularly in Europe, where peak gas prices have dramatically increased bills. This study defines fairness as a reform measure that carries both distributional and transitional effects on different income users and utilities. A flexible fairness-driven TOU tariff design is proposed, which (1) assigns households to different tariffs by income and estimated price responsiveness calculated from a long short-term memory (LSTM)-based counterfactual electricity consumption model and (2) filters potential tariffs to simultaneously lower total costs, distribute cost savings progressively, and maintain constant utility profitability. Using the 2009-10 Irish Smart Metering Trial based on more than 2000 users, under the proposed frameworks (b) and (c) with fairness constraints, no lower-income users have been observed to pay a higher bill than the higher-income users on average. Specifically, lower-income groups can achieve 8% and 10% bill reduction, respectively, under framework (b) partial bill protection and framework (c) full bill protection for the lower-income users. Our fairness-based TOU framework has successfully achieved fairness of distribution and transition, paving the way for a more equitable and sustainable electricity system.

Stochastic First-Order Algorithms for Constrained Distributionally Robust Optimization

We consider distributionally robust optimization (DRO) problems, reformulated as distributionally robust feasibility (DRF) problems, with multiple expectation constraints. We propose a generic stochastic first-order meta-algorithm, where the decision variables and uncertain distribution parameters are each updated separately by applying stochastic first-order methods. We then specialize our results to the case of using two specific versions of stochastic mirror descent (SMD): (i) a novel approximate version of SMD to update the decision variables, and (ii) the bandit mirror descent method to update the distribution parameters in the case of [Formula: see text]-divergence sets. For this specialization, we demonstrate that the total number of iterations is independent of the dimensions of the decision variables and distribution parameters. Moreover, the cost per iteration to update both sets of variables is nearly independent of the dimension of the distribution parameters, allowing for high-dimensional ambiguity sets. Furthermore, we show that the total number of iterations of our algorithm has a logarithmic dependence on the number of constraints. Experiments on logistic regression with fairness constraints, personalized parameter selection in a social network, and the multi-item newsvendor problem verify the theoretical results and show the usefulness of the algorithm, in particular when the dimension of the distribution parameters is large. History: Accepted by Antonio Frangioni, Area Editor for Design & Analysis of Algorithms—Continuous. Funding: This work was supported by the National Science Foundation [Grant 2112533]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/ijoc.2023.0167 .

A Personalized Framework for Consumer and Producer Group Fairness Optimization in Recommender Systems

In recent years, there has been an increasing recognition that when machine learning (ML) algorithms are used to automate decisions, they may mistreat individuals or groups, with legal, ethical, or economic implications. Recommender systems are prominent examples of these ML systems that aid users in making decisions. The majority of past literature research on recommender systems fairness treats user and item fairness concerns independently, ignoring the fact that recommender systems function in a two-sided marketplace. In this article, we propose CP-FairRank , an optimization-based re-ranking algorithm that seamlessly integrates fairness constraints from both the consumer and producer side in a joint objective framework. The framework is generalizable and may take into account varied fairness settings based on group segmentation, recommendation model selection, and domain, which is one of its key characteristics. For instance, we demonstrate that the system may jointly increase consumer and producer fairness when (un)protected consumer groups are defined on the basis of their activity level and main-streamness , while producer groups are defined according to their popularity level. For empirical validation, through large-scale on eight datasets and four mainstream collaborative filtering recommendation models, we demonstrate that our proposed strategy is able to improve both consumer and producer fairness without compromising or very little overall recommendation quality, demonstrating the role algorithms may play in avoiding data biases. Our results on different group segmentation also indicate that the amount of improvement can vary and is dependent on group segmentation, indicating that the amount of bias produced and how much the algorithm can improve it depend on the protected group definition, a factor that, to our knowledge, has not been examined in great depth in previous studies but rather is highlighted by the results discovered in this study.

Modification-fair cluster editing

The classic Cluster Editing problem (also known as Correlation Clustering) asks to transform a given graph into a disjoint union of cliques (clusters) by a small number of edge modifications. When applied to vertex-colored graphs (the colors representing subgroups), standard algorithms for the NP-hard Cluster Editing problem may yield solutions that are biased towards subgroups of data (e.g., demographic groups), measured in the number of modifications incident to the members of the subgroups. We propose a modification fairness constraint which ensures that the number of edits incident to each subgroup is proportional to its size. To start with, we study Modification-Fair Cluster Editing for graphs with two vertex colors. We show that the problem is NP-hard even if one may only insert edges within a subgroup; note that in the classic “non-fair” setting, this case is trivially polynomial-time solvable. However, in the more general editing form, the modification-fair variant remains fixed-parameter tractable with respect to the number of edge edits. We complement these and further theoretical results with an empirical analysis of our model on real-world social networks where we find that the price of modification-fairness is surprisingly low, that is, the cost of optimal modification-fair solutions differs from the cost of optimal “non-fair” solutions only by a small percentage.

Boosting Fair Classifier Generalization through Adaptive Priority Reweighing

With the increasing penetration of machine learning applications in critical decision-making areas, calls for algorithmic fairness are more prominent. Although there have been various modalities to improve algorithmic fairness through learning with fairness constraints, their performance does not generalize well in the test set. A performance-promising fair algorithm with better generalizability is needed. This paper proposes a novel adaptive reweighing method to eliminate the impact of the distribution shifts between training and test data on model generalizability. Most previous reweighing methods propose to assign a unified weight for each (sub)group. Rather, our method granularly models the distance from the sample predictions to the decision boundary. Our adaptive reweighing method prioritizes samples closer to the decision boundary and assigns a higher weight to improve the generalizability of fair classifiers. Extensive experiments are performed to validate the generalizability of our adaptive priority reweighing method for accuracy and fairness measures (i.e., equal opportunity, equalized odds, and demographic parity) in tabular benchmarks. We also highlight the performance of our method in improving the fairness of language and vision models. The code is available at https://github.com/che2198/APW .

Fair Pension Policies with Occupation-Specific Ageing

Abstract We study the optimal design of a fair public pension system in a multi-period overlapping generations model with occupation-specific morbidity and mortality that depends on the retirement age. The fairness constraint acts as institutional device ensuring that lifetime returns to contributions are equal across occupational groups. We consider group-specific replacement rates and a calculatory interest rate for early contributions as policy instruments. Calibrating the model to Germany, we find that the transition to optimal fair pension policies may induce early retirement of blue-collar workers and significantly raises their lifetime pension benefits and welfare. Aggregate welfare increases in all fair pension scenarios.

Wasserstein Robust Classification with Fairness Constraints

Problem definition: Data analytics models and machine learning algorithms are increasingly deployed to support consequential decision-making processes, from deciding which applicants will receive job offers and loans to university enrollments and medical interventions. However, recent studies show these models may unintentionally amplify human bias and yield significant unfavorable decisions to specific groups. Methodology/results: We propose a distributionally robust classification model with a fairness constraint that encourages the classifier to be fair in the equality of opportunity criterion. We use a type-[Formula: see text] Wasserstein ambiguity set centered at the empirical distribution to represent distributional uncertainty and derive a conservative reformulation for the worst-case equal opportunity unfairness measure. We show that the model is equivalent to a mixed binary conic optimization problem, which standard off-the-shelf solvers can solve. We propose a convex, hinge-loss-based model for large problem instances whose reformulation does not incur binary variables to improve scalability. Moreover, we also consider the distributionally robust learning problem with a generic ground transportation cost to hedge against the label and sensitive attribute uncertainties. We numerically examine the performance of our proposed models on five real-world data sets related to individual analysis. Compared with the state-of-the-art methods, our proposed approaches significantly improve fairness with negligible loss of predictive accuracy in the testing data set. Managerial implications: Our paper raises awareness that bias may arise when predictive models are used in service and operations. It generally comes from human bias, for example, imbalanced data collection or low sample sizes, and is further amplified by algorithms. Incorporating fairness constraints and the distributionally robust optimization (DRO) scheme is a powerful way to alleviate algorithmic biases. Funding: This work was supported by the National Science Foundation [Grants 2342505 and 2343869] and the Chinese University of Hong Kong [Grant 4055191]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/msom.2022.0230 .

Dynamic Environment Responsive Online Meta-Learning with Fairness Awareness

The fairness-aware online learning framework has emerged as a potent tool within the context of continuous lifelong learning. In this scenario, the learner’s objective is to progressively acquire new tasks as they arrive over time, while also guaranteeing statistical parity among various protected sub-populations, such as race and gender when it comes to the newly introduced tasks. A significant limitation of current approaches lies in their heavy reliance on the i.i.d (independent and identically distributed) assumption concerning data, leading to a static regret analysis of the framework. Nevertheless, it’s crucial to note that achieving low static regret does not necessarily translate to strong performance in dynamic environments characterized by tasks sampled from diverse distributions. In this article, to tackle the fairness-aware online learning challenge in evolving settings, we introduce a unique regret measure, FairSAR, by incorporating long-term fairness constraints into a strongly adapted loss regret framework. Moreover, to determine an optimal model parameter at each time step, we introduce an innovative adaptive fairness-aware online meta-learning algorithm, referred to as FairSAOML. This algorithm possesses the ability to adjust to dynamic environments by effectively managing bias control and model accuracy. The problem is framed as a bi-level convex-concave optimization, considering both the model’s primal and dual parameters, which pertain to its accuracy and fairness attributes, respectively. Theoretical analysis yields sub-linear upper bounds for both loss regret and the cumulative violation of fairness constraints. Our experimental evaluation of various real-world datasets in dynamic environments demonstrates that our proposed FairSAOML algorithm consistently outperforms alternative approaches rooted in the most advanced prior online learning methods.

FairSNA: Algorithmic Fairness in Social Network Analysis

In recent years, designing fairness-aware methods has received much attention in various domains, including machine learning, natural language processing, and information retrieval. However, in social network analysis (SNA), designing fairness-aware methods for various research problems by considering structural bias and inequalities of large-scale social networks has not received much attention. In this work, we highlight how the structural bias of social networks impacts the fairness of different SNA methods. We further discuss fairness aspects that should be considered while proposing network structure-based solutions for different SNA problems, such as link prediction, influence maximization, centrality ranking, and community detection. This survey-cum-vision clearly highlights that very few works have considered fairness and bias while proposing solutions; even these works are mainly focused on some research topics, such as link prediction, influence maximization, and PageRank. However, fairness has not yet been addressed for other research topics, such as influence blocking and community detection. We review the state of the art for different research topics in SNA, including the considered fairness constraints, their limitations, and our vision. This survey also covers evaluation metrics, available datasets and synthetic network generating models used in such studies. Finally, we highlight various open research directions that require researchers’ attention to bridge the gap between fairness and SNA.

Multi-User Redirected Walking in Separate Physical Spaces for Online VR Scenarios.

With the recent rise of Metaverse, online multiplayer VR applications are becoming increasingly prevalent worldwide. However, as multiple users are located in different physical environments, different reset frequencies and timings can lead to serious fairness issues for online collaborative/competitive VR applications. For the fairness of online VR apps/games, an ideal online RDW strategy must make the locomotion opportunities of different users equal, regardless of different physical environment layouts. The existing RDW methods lack the scheme to coordinate multiple users in different PEs, and thus have the issue of triggering too many resets for all the users under the locomotion fairness constraint. We propose a novel multi-user RDW method that is able to significantly reduce the overall reset number and give users a better immersive experience by providing a fair exploration. Our key idea is to first find out the "bottleneck" user that may cause all users to be reset and estimate the time to reset given the users' next targets, and then redirect all the users to favorable poses during that maximized bottleneck time to ensure the subsequent resets can be postponed as much as possible. More particularly, we develop methods to estimate the time of possibly encountering obstacles and the reachable area for a specific pose to enable the prediction of the next reset caused by any user. Our experiments and user study found that our method outperforms existing RDW methods in online VR applications.