Notion Of Fairness Research Articles

Dynamic fairness in repeated bargaining with risk

We study dynamic notions of fairness via an experiment of a two-round bilateral bargaining environment, where the payoff to one player is subject to ex-post risk, while the other player receives a fixed payment, effectively making the player exposed to risk a residual claimant. The ex-post risk not only provides substantive issues for bargaining parties to resolve in the experiment – i.e. what is a fair compensation for the exposure to risk – it also results in the endogenous formation of reference points for the second round due to ex post inequality after the realization of uncertainty in the first round. We find support for a “payback” hypothesis. That is, agreements in the second round significantly differ from the first round in a manner consistent with reducing the inequality that arose due to the initial pie realization.

Children’s consideration of collaboration and merit when making sharing decisions in private

Young children share equally when they acquire resources through collaboration with a partner, yet it is unclear whether they do so because in such contexts resources are encountered as common and distributed in front of the recipient or because collaboration promotes a sense of work-based fairness. In the current studies, 5- and 8-year-old children from Germany (N = 193) acquired resources either by working individually alongside or by collaborating with a peer. After finding out that the partner’s container was empty, they decided in private whether they wanted to donate some resources to the peer. When both partners had worked with equal efforts (Study 1), children shared more after collaboration than after individual work. When one partner had worked with much more effort than the other (Study 2), children shared more with a harder-working partner than with a less-working partner independently of whether they had collaborated or worked individually. Younger children were more generous than older children, in particular after collaboration. These findings support the view that collaboration promotes a genuine sense of fairness in young children, but they also indicate that merit-based notions of fairness in the context of work may develop independently of collaboration, at least by the beginning of middle childhood and in Western societies.

The role of political elites in income tax system design and tax fairness

Given the importance of fairness in the public acceptance of tax systems (Mirrlees et al., 2011), understanding the role of political elites in the tax design process and how this impacts fairness is important. In this paper, we use an interpretative approach to examine the design and ultimate rejection of the Kansas tax policy experiment that ran from May 2012 to June 2017. The central design element of this experiment was the fundamental disruption of the long-standing principle of horizontal equity (fairness), in which non-wage business income was relieved of income tax, while wage income was taxed. By interpreting this experiment through a Bourdieusian lens, we identify five key themes that are helpful in understanding income tax design, tax fairness, and the role played by political elites. Overall, we contribute to the literature by highlighting the way in which cultural capital may be used by political elites to reformulate groups of taxpayers in order to justify unequal treatment of equal taxpayers, which represents a violation of the notion of horizontal equity (fairness). We also highlight that although such violations of fairness in income tax system design may be sustainable in the short run, over the longer term, notions of fairness are so deeply embedded in the habitus of individual taxpayers that a return to the status quo is likely.

Exchange Networks with Stochastic Matching

This paper tries to prove that the outcomes stemming from interactions on assignment markets bring about coordination in case of a stochastic matching subject to various forms of expectations. We consider an exchange network with stochastic matching between the pairs of players and analyze the dynamics of bargaining in such a market. The cases of convergent expectations, divergent expectations and of social preferences are studied. The extension of earlier works lies in the consideration of a stochastic matching on a graph dependent on the weights of edges. The results show that, in all three cases, the dynamics converges rapidly to the generalized Nash bargaining solution, which is an equilibrium that combines notions of stability and fairness. In the first two scenarios, the numerical simulations reveal that the convergence toward a fixed point is speedily achieved at the value of the outside option. In the third scenario, the fixed point promptly converges to the value of the outside option supplemented by the surplus share.

Algorithmic Fairness and Statistical Discrimination

AbstractAlgorithmic fairness is a new interdisciplinary field of study focused on how to measure whether a process, or algorithm, may unintentionally produce unfair outcomes, as well as whether or how the potential unfairness of such processes can be mitigated. Statistical discrimination describes a set of informational issues that can induce rational (i.e., Bayesian) decision‐making to lead to unfair outcomes even in the absence of discriminatory intent. In this article, we provide overviews of these two related literatures and draw connections between them. The comparison illustrates both the conflict between rationality and fairness and the importance of endogeneity (e.g., “rational expectations” and “self‐fulfilling prophecies”) in defining and pursuing fairness. Taken in concert, we argue that the two traditions suggest a value for considering new fairness notions that explicitly account for how the individual characteristics an algorithm intends to measure may change in response to the algorithm.

Finite-trace and generalized-reactivity specifications in temporal synthesis

Linear Temporal Logic (LTL) synthesis aims at automatically synthesizing a program that complies with desired properties expressed in LTL. Unfortunately it has been proved to be too difficult computationally to perform full LTL synthesis. There have been two success stories with LTL synthesis, both having to do with the form of the specification. The first is the GR(1) approach: use safety conditions to determine the possible transitions in a game between the environment and the agent, plus one powerful notion of fairness, Generalized Reactivity(1), or GR(1). The second, inspired by AI planning, is focusing on finite-trace temporal synthesis, with LTLf\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$_f$$\\end{document} (LTL on finite traces) as the specification language. In this paper we take these two lines of work and bring them together. We first study the case in which we have an LTLf\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$_f$$\\end{document} agent goal and a GR(1) environment specification. We then add to the framework safety conditions for both the environment and the agent, obtaining a highly expressive yet still scalable form of LTL synthesis.

Optimal Joint Power and Rate Adaptation for Awareness and Congestion Control in Vehicular Networks

To support inter-vehicular applications, vehicles broadcast beacons with information about their state. Congestion may occur when the load on the channel due to beaconing can prevent the transmission of other types of messages. In this paper, we propose a distributed algorithm for optimal joint adaptation of transmit power and beaconing rate for congestion and awareness control. Our approach is based on a network utility formulation of the congestion control problem, which allows us to induce a desired fairness notion and set different priorities for vehicles. We formulate the general problem but, since it is not convex, we assume Rayleigh fading and derive an algorithm, called PRAIOS, that solves the optimization problem in a decentralized way with convergence guarantees. Our results, validated with realistic simulations in both static and dynamic scenarios and compared with other proposals, show that PRAIOS quickly converges to close to optimal allocations, while keeping the maximum beaconing load at the desired level. They also show that it can control the load when the fading is not Rayleigh. Applications can dynamically set their requirements as constraints, that are enforced by the algorithm while complying with the maximum load, which allows seamless integration of operational requirements into the control framework.

Learning to Design Fair and Private Voting Rules

Voting is used widely to identify a collective decision for a group of agents, based on their preferences. In this paper, we focus on evaluating and designing voting rules that support both the privacy of the voting agents and a notion of fairness over such agents. To do this, we introduce a novel notion of group fairness and adopt the existing notion of local differential privacy. We then evaluate the level of group fairness in several existing voting rules, as well as the trade-offs between fairness and privacy, showing that it is not possible to always obtain maximal economic efficiency with high fairness or high privacy levels. Then, we present both a machine learning and a constrained optimization approach to design new voting rules that are fair while maintaining a high level of economic efficiency. Finally, we empirically examine the effect of adding noise to create local differentially private voting rules and discuss the three-way trade-off between economic efficiency, fairness, and privacy. This paper appears in the special track on AI & Society.

Predictability and fairness in load aggregation and operations of virtual power plants

In power systems, one wishes to regulate the aggregate demand of an ensemble of distributed energy resources (DERs), such as controllable loads and battery energy storage systems. We propose a notion of predictability and fairness, which suggests that the long-term averages of prices or incentives offered should be independent of the initial states of the operators of the DER, the aggregator, and the power grid. We show that this notion cannot be guaranteed with many traditional controllers used by the load aggregator, including the usual proportional–integral (PI) controller. However, even considering the non-linearity of the alternating-current model, this notion of predictability and fairness can be guaranteed for incrementally input-to-state stable (iISS) controllers, under mild assumptions.

Fair and efficient allocation with few agent types, few item types, or small value levels

In fair division of indivisible goods, allocations that satisfy fairness and efficiency simultaneously are highly desired but may not exist or, even if they do exist, are computationally hard to find. Conditions under which such allocations, or allocations satisfying specific levels of fairness and efficiency simultaneously, can be efficiently found have thus been explored. Following this line of research, this study is concerned with the problem in a high-multiplicity setting where instances come with certain parameters, including agent types, item types, and value levels. Particularly, we address two computational problems. First, we wish to compute fair and Pareto-optimal allocations, w.r.t. any of the common fairness criteria: proportionality, maximin share, and max-min fairness. Second, we seek to find a max-min fair allocation that is efficient in the sense of maximizing utilitarian social welfare. We show that the first problem is tractable for most of the fairness criteria when the number of item types is fixed, or when at least two of the three parameters are fixed. For the second problem, we model it as a bi-criteria optimization problem that is solved approximately by determining an approximate Pareto set of bounded size. Our results are obtained based on dynamic programming and linear programming approaches. Our techniques strengthen known methods and can be potentially applied to other notions of fairness and efficiency.

Thou shalt covet the average of thy neighbors' cakes

We prove an Ω(n2) lower bound on the query complexity of local proportionality in the Robertson-Webb cake-cutting model. Local proportionality requires that each agent prefer their allocation to the average of their neighbors' allocations in some undirected social network. It is a weaker fairness notion than envy-freeness, which also has query complexity Ω(n2), and generally incomparable to proportionality, which has query complexity Θ(nlog⁡n). This result separates the complexity of local proportionality from that of ordinary proportionality, confirming the intuition that finding a locally proportional allocation is a more difficult computational problem.

"I'm not sure what difference is between their content and mine, other than the person itself"

How social media platforms could fairly conduct content moderation is gaining attention from society at large. Researchers from HCI and CSCW have investigated whether certain factors could affect how users perceive moderation decisions as fair or unfair. However, little attention has been paid to unpacking or elaborating on the formation processes of users' perceived (un)fairness from their moderation experiences, especially users who monetize their content. By interviewing 21 for-profit YouTubers (i.e., video content creators), we found three primary ways through which participants assess moderation fairness, including equality across their peers, consistency across moderation decisions and policies, and their voice in algorithmic visibility decision-making processes. Building upon the findings, we discuss how our participants' fairness perceptions demonstrate a multi-dimensional notion of moderation fairness and how YouTube implements an algorithmic assemblage to moderate YouTubers. We derive translatable design considerations for a fairer moderation system on platforms affording creator monetization.

A Comparative Study of Intentional Homicide (al-qatl al ʿamd) in Sharīʿah and the Modern Human Rights Concept of Right to Life for Murderers

This study sets forth the classical positions on the question of al-qatl al- ʿamd (intentional homicide), its definition, its scope and the application of the relevant punishment. The extent of compatibility, if any, with the modern human rights concept of right to life for murderers will also be explored. The latter concept is considered alongside the relevant provision of the Bill of Rights contained in the South African Constitution, including relevant case law for consideration to be applied in a modern context. This article advocates the idea that what is suitable for a specific country will be a function of its history, unique culture, and socio-political and legal conventions based on historical experiences and development. It also sets forth what is currently exercised in certain contexts, and whether the concept of al-qatl al- ‘amd would imperil the interests of the community. This study also tracks the ways in which leniency and avenues for repentance, recompense and forgiveness (al- ʿafwu) are navigated around social expediencies including Islamic notions of justice and fairness and what is appropriate in a modern context.

A unifying and general account of fairness measurement in recommender systems

Fairness is fundamental to all information access systems, including recommender systems. However, the landscape of fairness definition and measurement is quite scattered with many competing definitions that are partial and often incompatible. There is much work focusing on specific – and different – notions of fairness and there exist dozens of metrics of fairness in the literature, many of them redundant and most of them incompatible. In contrast, to our knowledge, there is no formal framework that covers all possible variants of fairness and allows developers to choose the most appropriate variant depending on the particular scenario. In this paper, we aim to define a general, flexible, and parameterizable framework that covers a whole range of fairness evaluation possibilities. Instead of modeling the metrics based on an abstract definition of fairness, the distinctive feature of this study compared to the current state of the art is that we start from the metrics applied in the literature to obtain a unified model by generalization. The framework is grounded on a general work hypothesis: interpreting the space of users and items as a probabilistic sample space, two fundamental measures in information theory (Kullback–Leibler Divergence and Mutual Information) can capture the majority of possible scenarios for measuring fairness on recommender system outputs. In addition, earlier research on fairness in recommender systems could be viewed as single-sided, trying to optimize some form of equity across either user groups or provider/procurer groups, without considering the user/item space in conjunction, thereby overlooking/disregarding the interplay between user and item groups. Instead, our framework includes the notion of statistical independence between user and item groups. We finally validate our approach experimentally on both synthetic and real data according to a wide range of state-of-the-art recommendation algorithms and real-world data sets, showing that with our framework we can measure fairness in a general, uniform, and meaningful way.

How institutional logics shape fairness in crowdsourcing: The case of Threadless

Fairness is essential for successful crowdsourcing. Without it, companies run the risk of consumers not participating, or worse, sabotaging the crowdsourcing initiative. Yet little is known about how consumers determine what is fair in crowdsourcing. Building on theories of organizational justice and institutional logics, and using a longitudinal netnography of Threadless, a popular crowdsourcing platform, this paper shows how experiences of fairness stem from the interaction between two conflicting crowdsourcing logics: the logic of renewal and the logic of community. The two logics inform notions of fairness in crowdsourcing contests across procedural, distributive, and interactional justice dimensions. A balance between the two logics is ideal for maintaining fairness among a crowdsourcing community. We show the conditions in which crowdsourcing participants tolerate transgressions to each justice dimension, consequently emphasizing one logic over the other. Overall, our study advances theory on crowdsourcing logics and how they guide notions of procedural, distributive, and interactional fairness in crowdsourcing. Our study also offers new guidance on how to manage fairness in crowdsourcing.

Portioning using ordinal preferences: Fairness and efficiency

A divisible public resource is to be divided among projects. We study rules that decide on a distribution of the budget when voters have ordinal preference rankings over projects. Examples of such portioning problems are participatory budgeting, time shares, and parliament elections. We introduce a family of rules for portioning, inspired by positional scoring rules. Rules in this family are given by a scoring vector (such as plurality or Borda) associating a positive value with each rank in a vote, and an aggregation function such as leximin or the Nash product. Our family contains well-studied rules, but most are new. We discuss computational and normative properties of our rules. We focus on fairness, and introduce the SD-core, a group fairness notion. Our Nash rules are in the SD-core, and the leximin rules satisfy individual fairness properties. Both are Pareto-efficient.

Equitable identification of gifted students with the relationship of religiosity and ethical sensitivity level of teachers

As teachers are the gatekeepers in identification procedure, it is crucial to understand how teachers’ openness and beliefs may affect the equitable identification. To be specific, we focused on how teachers’ level of religious beliefs and ethical sensitivity impact notions of fairness in the identification of gifted students, particularly those from traditionally underrepresented groups. Study participants were 46 teachers who attended a Midwestern Catholic University in the U.S. Based on the results measuring teachers’ religiosity, ethical sensitivity, and equitable identification index from vignettes, we found that there were no statistically significant score differences of religiosity and ethics level of teachers when they were divided into two groups, one showing equitable identification versus the other comparatively not. However, the majority of teachers (70%) fairly identified students regardless of their religiosity, ethical sensitivity, school experiences, confidence level serving gifted students, and current academic status when the vignettes had different students’ race and family income status information.

Computing welfare-Maximizing fair allocations of indivisible goods

We analyze the run-time complexity of computing allocations that are both fair and maximize the utilitarian social welfare, defined as the sum of agents’ utilities. We focus on two tractable fairness concepts: envy-freeness up to one item (EF1) and proportionality up to one item (PROP1). We consider two computational problems: (1) Among the utilitarian-maximal allocations, decide whether there exists one that is also fair; (2) among the fair allocations, compute one that maximizes the utilitarian welfare. We show that both problems are strongly NP-hard when the number of agents is variable, and remain NP-hard for a fixed number of agents greater than two. For the special case of two agents, we find that problem (1) is polynomial-time solvable, while problem (2) remains NP-hard. Finally, with a fixed number of agents, we design pseudopolynomial-time algorithms for both problems. We extend our results to the stronger fairness notions envy-freeness up to any item (EFx) and proportionality up to any item (PROPx).

From fair predictions to just decisions? Conceptualizing algorithmic fairness and distributive justice in the context of data-driven decision-making.

Prediction algorithms are regularly used to support and automate high-stakes policy decisions about the allocation of scarce public resources. However, data-driven decision-making raises problems of algorithmic fairness and justice. So far, fairness and justice are frequently conflated, with the consequence that distributive justice concerns are not addressed explicitly. In this paper, we approach this issue by distinguishing (a) fairness as a property of the algorithm used for the prediction task from (b) justice as a property of the allocation principle used for the decision task in data-driven decision-making. The distinction highlights the different logic underlying concerns about fairness and justice and permits a more systematic investigation of the interrelations between the two concepts. We propose a new notion of algorithmic fairness called error fairness which requires prediction errors to not differ systematically across individuals. Drawing on sociological and philosophical discourse on local justice, we present a principled way to include distributive justice concerns into data-driven decision-making. We propose that allocation principles are just if they adhere to well-justified distributive justice principles. Moving beyond the one-sided focus on algorithmic fairness, we thereby make a first step toward the explicit implementation of distributive justice into data-driven decision-making.

FFTree: A flexible tree to handle multiple fairness criteria

The demand for transparency and fairness in AI-based decision-making systems is constantly growing. Organisations need to be assured that their applications, based on these technologies, behave fairly, without introducing negative social implications in relation to sensitive attributes such as gender or race. Since the notion of fairness is context dependent and not uniquely defined, studies in the literature have proposed various formalisation. In this work, we propose a novel, flexible, discrimination-aware decision-tree that allows the user to employ different fairness criteria depending on the application domain. Our approach enhances decision-tree classifiers to provide transparent and fair rules to final users.