Fairness Constraints Research Articles

When Fairness Meets Consistency in AHP Pairwise Comparisons

We propose introducing fairness constraints to one of the most famous multi-criteria decision-making methods, the analytic hierarchy process (AHP). We offer a solution that guarantees consistency while respecting legally binding fairness constraints in AHP pairwise comparison matrices. Through a synthetic experiment, we generate the comparison matrices of different sizes and ranges/levels of the initial parameters (i.e., consistency ratio and disparate impact). We optimize disparate impact for various combinations of these initial parameters and observed matrix sizes while respecting an acceptable level of consistency and minimizing deviations of pairwise comparison matrices (or their upper triangles) before and after the optimization. We use a metaheuristic genetic algorithm to set the dually motivating problem and operate a discrete optimization procedure (in connection with Saaty’s 9-point scale). The results confirm the initial hypothesis (with 99.5% validity concerning 2800 optimization runs) that achieving fair ranking while respecting consistency in AHP pairwise comparison matrices (when comparing alternatives regarding given criterium) is possible, thus meeting two challenging goals simultaneously. This research contributes to the initiatives directed toward unbiased decision-making, either automated or algorithm-assisted (which is the case covered by this research).

Social norm bias: residual harms of fairness-aware algorithms

Many modern machine learning algorithms mitigate bias by enforcing fairness constraints across coarsely-defined groups related to a sensitive attribute like gender or race. However, these algorithms seldom account for within-group heterogeneity and biases that may disproportionately affect some members of a group. In this work, we characterize Social Norm Bias (SNoB), a subtle but consequential type of algorithmic discrimination that may be exhibited by machine learning models, even when these systems achieve group fairness objectives. We study this issue through the lens of gender bias in occupation classification. We quantify SNoB by measuring how an algorithm’s predictions are associated with conformity to inferred gender norms. When predicting if an individual belongs to a male-dominated occupation, this framework reveals that “fair” classifiers still favor biographies written in ways that align with inferred masculine norms. We compare SNoB across algorithmic fairness techniques and show that it is frequently a residual bias, and post-processing approaches do not mitigate this type of bias at all.

Complexity and Monitoring of Economic Operations Using a Game-Theoretic Model for Cloud Computing

In this study, a model is presented for allocating cloud computing resources based on economic considerations using tools from game theory. The model, called the Non-Cooperative Game Resource Allocation Algorithm (NCGRAA), is designed to achieve the optimum stage in cloud computing. In addition, the Bargaining Game Resource Allocation Algorithm (BGRAA) is introduced to the existing system to develop the billing process within the constraints of availability and fairness. This system-based algorithm implements methods for converging on and improving the Nash Equilibrium and Nash Bargaining solutions. While the Nash equilibrium helps to develop decision-making concepts with game theory, one of its main goals is to achieve the desired outcome and avoid deviation from the working stage. Nash Bargaining is a unique solution that occurs between two parties and takes into account the process of bargaining to provide a fair solution that is scale invariant and independent. In recent years, cloud computing has become a popular way to manage computing services and enable producers and consumers to interact. This process allows users to obtain goods at an affordable cost from sellers according to their expectations. This research investigates the economic operation monitoring of cloud computing using the gaming theory model. A Static Negotiation Analysis Method with a Bargaining Process (SNAM-BP) for a dynamic conceptual framework is presented to display the weighted relationship between primary issues and keywords used to evaluate the potential partnership of each country.

Fairness in optimal operation of transactive smart networked modern multi-carrier energy systems: A two-stage optimization approach

Multi-Carrier Energy Systems (MCESs) have attracted global appeal and indicated broad outlooks in developing green energy resources in the community. The transactive networked operation of multiple MCESs improves economic opportunities and system flexibility. Nevertheless, how to fairly and convincingly allocate the total expenses among MCESs is a crucial challenge that is addressed in this paper. To tackle the fairness issue, a two-stage optimization framework in the transactive networked MCESs is developed. In the first stage, named the islanding operation mode, the operation cost of each MCES is minimized and the optimization outcomes are sent to the central operator of the transactive networked MCESs. In the second stage, the central operator solves an optimization problem to minimize the total operation cost of the transactive networked MCESs considering energy transactions among MCESs and fairness constraints. The fair cost allocation constraints lead to similar or less contributions in total operation cost in the transactive networked structure compared with the islanding mode. The proposed model is structured as a Mixed-Integer Linear Programming problem and formulated in GAMS modeling language. The results indicate that the proposed two-stage optimization approach decreases the operation cost of the networked MCESs by 16.7% while the fairness issue was satisfied.

Online Passenger Flow Control in Metro Lines

Crowds management during peak commuting hours is a key challenge facing metro systems worldwide, which results in serious safety concerns and unfair public transit service for commuters on different origin-destination (o-d) pairs. In “Online Passenger Flow Control in Metro Lines,” the authors investigate the impact of online decision making on the value of passenger flow control solution methodologies. The authors formulate the problem as a stochastic dynamic program with a fairness (fill rate) constraint and exploit Blackwell's approachability theorem and Fenchel duality to characterize the attainable service level of each o-d pair. They use these insights to develop online policies that can enable more passengers boarding a train (efficiency) as well as ensure equitable service level (fairness) provided to each o-d pair. Numerical experiments on a set of transit data from Beijing show that this approach performs well compared with existing benchmarks in the literature.

A back adjustment based dependent task offloading scheduling algorithm with fairness constraints in VEC networks

In Vehicular Edge Computing (VEC) networks, task offloading scheduling has been drawing more and more attention as an effective way to relieve the computational burden of vehicles. However, with the intelligent and networked development of vehicles, the complex data dependency between in-vehicle tasks brings challenges to offloading scheduling. Moreover, scheduling fairness has a growing impact on the average Quality of Service (QoS) of vehicles in the network. To this end, we propose a dependent task offloading scheduling algorithm with fairness constraints based on a back adjustment mechanism. First, to solve the execution constraint problem caused by dependent tasks and the scheduling fairness problem in multi-user scenarios, a two-level task sorting algorithm is given to determine the scheduling sequence of tasks. Then, the sequential task offloading scheduling process is modeled as a Markov Decision Process (MDP) and solved by a reinforcement learning method. Finally, a back adjustment mechanism is designed to resort the task sequence and achieve the required scheduling fairness by iterative process. The simulation results show that the proposed algorithm significantly improves the scheduling fairness and reduces the average application completion time compared with other algorithms.

A Multi-Objective Optimization Framework for Multi-Stakeholder Fairness-Aware Recommendation

Nowadays, most online services are hosted on multi-stakeholder marketplaces, where consumers and producers may have different objectives. Conventional recommendation systems, however, mainly focus on maximizing consumers’ satisfaction by recommending the most relevant items to each individual. This may result in unfair exposure of items, thus jeopardizing producer benefits. Additionally, they do not care whether consumers from diverse demographic groups are equally satisfied. To address these limitations, we propose a multi-objective optimization framework for fairness-aware recommendation, Multi-FR , that adaptively balances accuracy and fairness for various stakeholders with Pareto optimality guarantee. We first propose four fairness constraints on consumers and producers. In order to train the whole framework in an end-to-end way, we utilize the smooth rank and stochastic ranking policy to make these fairness criteria differentiable and friendly to back-propagation. Then, we adopt the multiple gradient descent algorithm to generate a Pareto set of solutions, from which the most appropriate one is selected by the Least Misery Strategy. The experimental results demonstrate that Multi-FR largely improves recommendation fairness on multiple stakeholders over the state-of-the-art approaches while maintaining almost the same recommendation accuracy. The training efficiency study confirms our model’s ability to simultaneously optimize different fairness constraints for many stakeholders efficiently.

On the task assignment with group fairness for spatial crowdsourcing

Task assignment, the core problem of Spatial Crowdsourcing (SC), is often modeled as an optimization problem with multiple constraints, and the quality and efficiency of its solution determines how well the SC system works. Fairness is a critical aspect of task assignment that affects worker participation and satisfaction. Although the existing studies on SC have noticed the fairness problem, they mainly focus on fairness at the individual level rather than at the group level. However, differences among groups in certain attributes (e.g. race, gender, age) can easily lead to discrimination in task assignment, which triggers ethical issues and even deteriorates the quality of the SC system. Therefore, we study the problem of task assignment with group fairness for SC. According to the principle of fair budget allocation, we define a well-designed constraint that can be considered in the task assignment problem of SC systems, resulting in assignment with group fairness. We mainly consider the task assignment problem in a common One-to-One SC system (O2-SC), and our goal is to maximize the quality of the task assignment while satisfying group fairness and other constraints such as budget and spatial constraints. Specifically, we first give the formal definition of task assignment with group fairness constraint for O2-SC. Then, we prove that it is essentially an NP-hard combinatorial optimization problem. Next, we provide a novel fast algorithm with theoretical guarantees to solve it. Finally, we conduct extensive experiments using both synthetic and real datasets. The experimental results show that the proposed constraint can significantly improve the group fairness level of algorithms, even for a completely random algorithm. The results also show that our algorithm can efficiently and effectively complete the task assignment of SC systems while ensuring group fairness.

Equitability and welfare maximization for allocating indivisible items

We study fair allocations of indivisible goods and chores in conjunction with system efficiency, measured by two social welfare functions, namely utilitarian and egalitarian welfare. To model preference, each agent is associated with a cardinal and additive valuation function. The fairness criteria we are concerned with are equitability up to any item (EQX) and equitability up to one item (EQ1). For the trade-off between fairness and efficiency, we investigate efficiency loss under these fairness constraints and establish the price of fairness. From the computational perspective, we provide a complete picture of the computational complexity of (i) deciding the existence of an EQX/EQ1 and welfare-maximizing allocation; (ii) computing a welfare maximizer among all EQX/EQ1 allocations.

Fairness constraint efficiency optimization for multiresource allocation in a cluster system serving internet of things

SummaryMassive, diverse, and high‐frequency Internet of Things (IoT) applications pose challenges to the operation of cluster systems that serve it. Fair and efficient multidimensional resource allocation is of great significance to the sustainable operation of these systems. However, most of the existing cluster multiresource allocation optimization researches focus too much on the fairness of resource allocation and ignore the efficiency. The unbalanced use of multidimensional system resources reduces the effective utilization of system resources, which seriously affects the service quality of IoT applications. In this paper, we define the multiresource fair and efficient sharing optimization as a fairness‐constrained efficiency optimization problem, which is from dynamics, discrete resources, and heterogeneous perspectives according to the characteristics of cluster system in practical. Moreover, we present a dynamic efficiency‐aware multiresource fair allocation algorithm, DEF, which can improve the ability of the cluster system to serve diverse IoT applications. In the algorithm, large jobs schedule to the servers that expect the least remaining resources. Simulations performed using Google cluster‐usage traces show that DEF can improve system resource utilization and guarantee the fairness of sharing among users.

Fairness-Oriented Transmission Schemes for Multiuser MISO Broadcast Channels

With the increasing number of Internet of Things (IoT), Industry 4.0 (I4.0), and mobile devices, it can be expected that base stations will have to serve more and more clients with a limited number of antennas. For their broadcast channels, nonorthogonal multiple access (NOMA) and blind interference alignment (BIA) are two efficient and commonly adopted transmission schemes. This paper conducts a comparison study on these techniques on a 3-user 2 × 1 multiple-input single-output (MISO) broadcast channel with a limited number of transmit antennas. Specifically, space-time block coding based NOMA (STBC-NOMA) and NOMA-assisted beamforming (NOMA-BF) are compared with BIA. Both perfect and imperfect successive interference cancellation (SIC) have been considered for NOMA-based schemes, and the theoretical achievable rates of all schemes have been derived. Furthermore, with a given fairness constraint among end users, the power allocation (PA) problems have been solved for cases when accurate channel state information is available at the transmitter (CSIT) as well as when only path loss information is available. Numerical results show the following: (1) none of the schemes under this study can always outperform the others under different SNR regions. (2) With imperfect SIC, NOMA-BF, and STBC-NOMA both suffer from a significant performance loss under a high SNR condition. (3) Fairness PA with only path loss information provides similar performance as that with perfect CSIT, thus partial CSIT is adequate for system or scheme designs in practice.

Influence Maximization under Fairness Budget Distribution in Online Social Networks

In social influence analysis, viral marketing, and other fields, the influence maximization problem is a fundamental one with critical applications and has attracted many researchers in the last decades. This problem asks to find a k-size seed set with the largest expected influence spread size. Our paper studies the problem of fairness budget distribution in influence maximization, aiming to find a seed set of size k fairly disseminated in target communities. Each community has certain lower and upper bounded budgets, and the number of each community’s elements is selected into a seed set holding these bounds. Nevertheless, resolving this problem encounters two main challenges: strongly influential seed sets might not adhere to the fairness constraint, and it is an NP-hard problem. To address these shortcomings, we propose three algorithms (FBIM1, FBIM2, and FBIM3). These algorithms combine an improved greedy strategy for selecting seeds to ensure maximum coverage with the fairness constraints by generating sampling through a Reverse Influence Sampling framework. Our algorithms provide a (1/2−ϵ)-approximation of the optimal solution, and require OkTlog(8+2ϵ)nln2δ+ln(kn)ϵ2, OkTlognϵ2k, and OTϵlogkϵlognϵ2k complexity, respectively. We conducted experiments on real social networks. The result shows that our proposed algorithms are highly scalable while satisfying theoretical assurances, and that the coverage ratios with respect to the target communities are larger than those of the state-of-the-art alternatives; there are even cases in which our algorithms reaches 100% coverage with respect to target communities. In addition, our algorithms are feasible and effective even in cases involving big data; in particular, the results of the algorithms guarantee fairness constraints.

Computation Bits Maximization in UAV-Assisted MEC Networks With Fairness Constraint

This article investigates an unmanned aerial vehicle (UAV)-assisted wireless-powered mobile-edge computing (MEC) system, where the UAV powers the mobile terminals by wireless power transfer (WPT) and provides computation service for them. We aim to maximize the computation bits of terminals while ensuring fairness among them. Considering the random trajectories of mobile terminals, we propose a soft actor–critic (SAC)-based UAV trajectory planning and resource allocation (SAC-TR) algorithm, which combines off-policy and maximum entropy reinforcement learning to improve the convergence of the algorithm. We design the reward as a heterogeneous function of computation bits, fairness, and destination. Simulation results show that SAC-TR can quickly adapt to varying network environments and outperform representative benchmarks in various situations.

When More Data Lead Us Astray: Active Data Acquisition in the Presence of Label Bias

An increased awareness concerning risks of algorithmic bias has driven a surge of efforts around bias mitigation strategies. A vast majority of the proposed approaches fall under one of two categories: (1) imposing algorithmic fairness constraints on predictive models, and (2) collecting additional training samples. Most recently and at the intersection of these two categories, methods that propose active learning under fairness constraints have been developed. However, proposed bias mitigation strategies typically overlook the bias presented in the observed labels. In this work, we study fairness considerations of active data collection strategies in the presence of label bias. We first present an overview of different types of label bias in the context of supervised learning systems. We then empirically show that, when overlooking label bias, collecting more data can aggravate bias, and imposing fairness constraints that rely on the observed labels in the data collection process may not address the problem. Our results illustrate the unintended consequences of deploying a model that attempts to mitigate a single type of bias while neglecting others, emphasizing the importance of explicitly differentiating between the types of bias that fairness-aware algorithms aim to address, and highlighting the risks of neglecting label bias during data collection.

A novel mixed-integer programming approach for the aircraft landing problem

Due to the upward trend in the aviation industry, the existing approaches for air traffic control need to be improved to achieve efficient schedules. This paper deals with the aircraft landing problem, which consists of determining a landing time for each aircraft within the radar range of an airport and allocating it to a runway. We propose an exact solution approach that involves mixed-integer linear programming. The objective is hereby to minimize the sum of weighted deviations from the target landing times under consideration of different safety, efficiency and fairness constraints. Despite of the problem’s NP-hardness, our method exhibits low execution times thanks to a modified modeling strategy and provides near-optimal results. Numerical experiments prove efficiency of the approach for different large airports.

Scheduling mobile dental clinics: A heuristic approach considering fairness among school districts.

Mobile dental clinics (MDCs) are suitable solutions for servicing people living in rural and urban areas that require dental healthcare. MDCs can provide dental care to the most vulnerable high-school students. However, scheduling MDCs to visit patients is critical to developing efficient dental programs. Here, we study a mobile dental clinic scheduling problem that arises from the real-life logistics management challenge faced by a school-based mobile dental care program in Southern Chile. This problem involves scheduling MDCs to treat high-school students at public schools while considering a fairness constraint among districts. Schools are circumscribed into districts, and by program regulations, at least 50% of the students in each district must receive dental care during the first semester. Fairness prevents some districts from waiting more time to receive dental care than others. We model the problem as a parallel machine scheduling problem with sequence-dependent setup costs and batch due dates and propose a mathematical model and a genetic algorithm-based solution to solve the problem. Our computational results demonstrate the effectiveness of our approaches in obtaining near-optimal solutions. Finally, dental program managers can use the methodologies presented in this work to schedule mobile dental clinics and improve their operations.

Fairness in graph-based semi-supervised learning

Machine learning is widely deployed in society, unleashing its power in a wide range of applications owing to the advent of big data. One emerging problem faced by machine learning is the discrimination from data, and such discrimination is reflected in the eventual decisions made by the algorithms. Recent study has proved that increasing the size of training (labeled) data will promote the fairness criteria with model performance being maintained. In this work, we aim to explore a more general case where quantities of unlabeled data are provided, indeed leading to a new form of learning paradigm, namely fair semi-supervised learning. Taking the popularity of graph-based approaches in semi-supervised learning, we study this problem both on conventional label propagation method and graph neural networks, where various fairness criteria can be flexibly integrated. Our developed algorithms are proved to be non-trivial extensions to the existing supervised models with fairness constraints. Extensive experiments on real-world datasets exhibit that our methods achieve a better trade-off between classification accuracy and fairness than the compared baselines.

Satisfying complex top- k fairness constraints by preference substitutions

Given m users (voters), where each user casts her preference for a single item (candidate) over n items (candidates) as a ballot, the preference aggregation problem returns k items (candidates) that have the k highest number of preferences (votes). Our work studies this problem considering complex fairness constraints that have to be satisfied via proportionate representations of different values of the group protected attribute(s) in the top- k results. Precisely, we study the margin finding problem under single ballot substitutions , where a single substitution amounts to removing a vote from candidate i and assigning it to candidate j and the goal is to minimize the number of single ballot substitutions needed to guarantee that the top-k results satisfy the fairness constraints. We study several variants of this problem considering how top- k fairness constraints are defined, (i) MFBinaryS and MFMultiS are defined when the fairness (proportionate representation) is defined over a single, binary or multivalued, protected attribute, respectively; (ii) MF-Multi2 is studied when top- k fairness is defined over two different protected attributes; (iii) MFMulti3+ investigates the margin finding problem, considering 3 or more protected attributes. We study these problems theoretically, and present a suite of algorithms with provable guarantees. We conduct rigorous large scale experiments involving multiple real world datasets by appropriately adapting multiple state-of-the-art solutions to demonstrate the effectiveness and scalability of our proposed methods.

Happiness maximizing sets under group fairness constraints

Finding a happiness maximizing set (HMS) from a database, i.e., selecting a small subset of tuples that preserves the best score with respect to any nonnegative linear utility function, is an important problem in multi-criteria decision-making. When an HMS is extracted from a set of individuals to assist data-driven algorithmic decisions such as hiring and admission, it is crucial to ensure that the HMS can fairly represent different groups of candidates without bias and discrimination. However, although the HMS problem was extensively studied in the database community, existing algorithms do not take group fairness into account and may provide solutions that under-represent some groups. In this paper, we propose and investigate a fair variant of HMS (FairHMS) that not only maximizes the minimum happiness ratio but also guarantees that the number of tuples chosen from each group falls within predefined lower and upper bounds. Similar to the vanilla HMS problem, we show that FairHMS is NP-hard in three and higher dimensions. Therefore, we first propose an exact interval cover-based algorithm called IntCov for FairHMS on two-dimensional databases. Then, we propose a bicriteria approximation algorithm called BiGreedy for FairHMS on multi-dimensional databases by transforming it into a submodular maximization problem under a matroid constraint. We also design an adaptive sampling strategy to improve the practical efficiency of BiGreedy. Extensive experiments on real-world and synthetic datasets confirm the efficacy and efficiency of our proposal.

Sequential Submodular Maximization and Applications to Ranking an Assortment of Products

The advents of online retailing and advertising have created new opportunities for online platforms to incorporate algorithmic techniques to improve shoppers’ experience and drive user engagement, which, in return, can help with the long-term growth of these platforms, and also to help with having socially aware operations that consider fairness across different demographic groups. Motivated by the product-ranking problem in online shopping, this paper introduces and studies a new class of combinatorial optimization problems over the space of permutations, which is referred to as “sequential submodular maximization.” Using this class of problems, it provides algorithmic solutions for maximizing users’ engagement and also for balancing the users’ engagement across different demographic groups of shoppers to obtain fairness. In particular, they propose an optimal (1 − 1/e)-approximation algorithms for maximizing users’ engagement and a bicriteria ((1 − 1/e)2, (1 − 1/e)2) for maximizing users’ engagement subject to group fairness constraints.