Notion Of Fairness Research Articles

The Price of Connectivity in Fair Division

We study the allocation of indivisible goods that form an undirected graph and quantify the loss of fairness when we impose a constraint that each agent must receive a connected subgraph. Our focus is on well-studied fairness notions including envy-freeness and maximin share fairness. We introduce the price of connectivity to capture the largest gap between the graph-specific and the unconstrained maximin share, and derive bounds on this quantity which are tight for large classes of graphs in the case of two agents and for paths and stars in the general case. For instance, with two agents we show that for biconnected graphs it is possible to obtain at least $3/4$ of the maximin share with connected allocations, while for the remaining graphs the guarantee is at most $1/2$. In addition, we determine the optimal relaxation of envy-freeness that can be obtained with each graph for two agents, and characterize the set of trees and complete bipartite graphs that always admit an allocation satisfying envy-freeness up to one good (EF1) for three agents. Our work demonstrates several applications of graph-theoretic tools and concepts to fair division problems.

Who Pays for Canada? Taxes and Fairness ed. by E. A. Heaman and David Tough

Reviewed by: Who Pays for Canada? Taxes and Fairness ed. by E. A. Heaman and David Tough Dimitry Anastakis Heaman, E. A. and David Tough eds. – Who Pays for Canada? Taxes and Fairness. Montréal and Kingston: McGill-Queen's University Press, 2020. 424 p. A few decades from now, historians may look back upon the "taxation turn" of the 2010s as an important historiographical interjection within Canada's evolving new political history. The troika of monographs produced by Elsbeth Heaman, David Tough (co-editors of this volume), and Shirley Tillotson (a contributor), have effectively reexamined and reframed how we think about taxation, the relationship between individuals and the state, and the very notion of fairness in society. Collectively, and whether they realize it or not, they have created a "taxation school" subfield, one which has sparked a vibrant historical debate that has deep implications for how we understand many of the questions that drive Canadian political economy today. Who Pays for Canada: Taxes and Fairness is a welcome addition to this newly established taxation canon. With five sections and 17 chapters that cover topics in comparative politics, history, economics, gender, "the making of the modern taxpayer," and "democratic tax accountability," the collection boasts a breadth of scholarly approaches that includes historians, philosophers, geographers, economists, political scientists, and business and law school professors. Despite this wide range of contributors and perspectives, the interdisciplinarity and accessibility of this volume is a strength. It aims at "setting new standards for interdisciplinarity and town-gown conversations" in a manner that avoids "insider baseball" (p. 19). And the collection largely succeeds. The preface, by former Parliamentary Budget Officer Kevin Page, and the introduction, "Broadening the Tax Conversation," set the tone and convincingly make their case for the book's interdisciplinary approach. In places, Who Pays for Canada? reads less like a traditional historical edited collection, and more like a special issue on taxation from the pages of Canadian Public Policy, or some such economics- or policy-driven publication. That's a good thing, and provides a refreshing, and frustratingly rare, expansion in the ambit of historians. Bridging the historical with the applied, as has been done in the collection, could stand to be pushed further outside of the academy, too. Wouldn't the Bank of Canada and Canada's ministries of finance be better served by having a few more historians or other types of scholars in their number, and a few less economists? This book's expansive range of convincing chapters only adds to any conviction in that regard, and it provides a great example of activist history in the policy realm. This activism is reflected in the book's two main currents. Many of the chapters examine ideas of taxation, forcing us to think about taxation as a conceptual notion, one that shapes, and is shaped by, history, comparative politics (across countries, and provinces), party politics, emotion (jealousy), knowledge, perception, and, above all, questions of fairness. Re-thinking taxation reveals the intellectual, psychological, and emotional shape of taxation, and how the humanity of taxation policy has been used (or abused) to influence who pays, who does not, and why that matters. At the same time, the book provides a refreshing connection to the material realities of ordinary Canadians, and how taxation shapes everyday existence. [End Page 200] Chapters that explore the taxation implications upon schooling, municipal government, families, gender, region, and colonialism provide insights into how tax regimes impact people, classes of people, and places, and how those impacts are uneven and unevenly distributed in society. This current, too, is about fairness, and how the policies and politics of taxation can shape the most fundamental notions of governance, identity and, of course, inequality. Ultimately, both currents, and all the chapters in the collection, remind us that history can show us that there are alternative paths and different ways of thinking about and implementing taxation policy to make it fairer, less unequal, and better for all Canadians. In that regard, this book, and the historiographical interjection from which it springs, should not only be read by historians, but all humanists and social scientists who are engaged with the work of creating...

The Causal Fairness Field Guide: Perspectives From Social and Formal Sciences.

Over the past several years, multiple different methods to measure the causal fairness of machine learning models have been proposed. However, despite the growing number of publications and implementations, there is still a critical lack of literature that explains the interplay of causality-based fairness notions with the social sciences of philosophy, sociology, and law. We hope to remedy this issue by accumulating and expounding upon the thoughts and discussions of causality-based fairness notions produced by both social and formal (specifically machine learning) sciences in this field guide. In addition to giving the mathematical backgrounds of several popular causality-based fair machine learning notions, we explain their connection to and interplay with the fields of philosophy and law. Further, we explore several criticisms of the current approaches to causality-based fair machine learning from a sociological viewpoint as well as from a technical standpoint. It is our hope that this field guide will help fair machine learning practitioners better understand how their causality-based fairness notions align with important humanistic values (such as fairness) and how we can, as a field, design methods and metrics to better serve oppressed and marginalized populaces.

Algorithmic fairness through group parities? The case of COMPAS-SAPMOC

Machine learning classifiers are increasingly used to inform, or even make, decisions significantly affecting human lives. Fairness concerns have spawned a number of contributions aimed at both identifying and addressing unfairness in algorithmic decision-making. This paper critically discusses the adoption of group-parity criteria (e.g., demographic parity, equality of opportunity, treatment equality) as fairness standards. To this end, we evaluate the use of machine learning methods relative to different steps of the decision-making process: assigning a predictive score, linking a classification to the score, and adopting decisions based on the classification. Throughout our inquiry we use the COMPAS system, complemented by a radical simplification of it (our SAPMOC I and SAPMOC II models), as our running examples. Through these examples, we show how a system that is equally accurate for different groups may fail to comply with group-parity standards, owing to different base rates in the population. We discuss the general properties of the statistics determining the satisfaction of group-parity criteria and levels of accuracy. Using the distinction between scoring, classifying, and deciding, we argue that equalisation of classifications/decisions between groups can be achieved thorough group-dependent thresholding. We discuss contexts in which this approach may be meaningful and useful in pursuing policy objectives. We claim that the implementation of group-parity standards should be left to competent human decision-makers, under appropriate scrutiny, since it involves discretionary value-based political choices. Accordingly, predictive systems should be designed in such a way that relevant policy goals can be transparently implemented. Our paper presents three main contributions: (1) it addresses a complex predictive system through the lens of simplified toy models; (2) it argues for selective policy interventions on the different steps of automated decision-making; (3) it points to the limited significance of statistical notions of fairness to achieve social goals.

How do the existing fairness metrics and unfairness mitigation algorithms contribute to ethical learning analytics?

AbstractWith the widespread use of learning analytics (LA), ethical concerns about fairness have been raised. Research shows that LA models may be biased against students of certain demographic subgroups. Although fairness has gained significant attention in the broader machine learning (ML) community in the last decade, it is only recently that attention has been paid to fairness in LA. Furthermore, the decision on which unfairness mitigation algorithm or metric to use in a particular context remains largely unknown. On this premise, we performed a comparative evaluation of some selected unfairness mitigation algorithms regarded in the fair ML community to have shown promising results. Using a 3‐year program dropout data from an Australian university, we comparatively evaluated how the unfairness mitigation algorithms contribute to ethical LA by testing for some hypotheses across fairness and performance metrics. Interestingly, our results show how data bias does not always necessarily result in predictive bias. Perhaps not surprisingly, our test for fairness‐utility tradeoff shows how ensuring fairness does not always lead to drop in utility. Indeed, our results show that ensuring fairness might lead to enhanced utility under specific circumstances. Our findings may to some extent, guide fairness algorithm and metric selection for a given context. Practitioner notesWhat is already known about this topic LA is increasingly being used to leverage actionable insights about students and drive student success. LA models have been found to make discriminatory decisions against certain student demographic subgroups—therefore, raising ethical concerns. Fairness in education is nascent. Only a few works have examined fairness in LA and consequently followed up with ensuring fair LA models. What this paper adds A juxtaposition of unfairness mitigation algorithms across the entire LA pipeline showing how they compare and how each of them contributes to fair LA. Ensuring ethical LA does not always lead to a dip in performance. Sometimes, it actually improves performance as well. Fairness in LA has only focused on some form of outcome equality, however equality of outcome may be possible only when the playing field is levelled. Implications for practice and/or policy Based on desired notion of fairness and which segment of the LA pipeline is accessible, a fairness‐minded decision maker may be able to decide which algorithm to use in order to achieve their ethical goals. LA practitioners can carefully aim for more ethical LA models without trading significant utility by selecting algorithms that find the right balance between the two objectives. Fairness enhancing technologies should be cautiously used as guides—not final decision makers. Human domain experts must be kept in the loop to handle the dynamics of transcending fair LA beyond equality to equitable LA.

‘Rorting the System’: Police Detectives, Diversity, and Workplace Advantage

Internal workplace practices and policies in policing are based on a notion of fairness and equal opportunity. Yet police organizations are frequently criticized for discriminatory policing practices, unfair and biased workplace practices, and poor interpersonal treatment of officers. Whilst there is a wide body of research examining diversity in relation to external police practices, there is a lack of knowledge regarding diversity and internal workplace practices; particularly from the perspective of police detectives who often have more substantial policing experience and longer employment histories than other non-commissioned officers. Contributing new findings to the extant policing literature, this research analyzes data collected from interviews with twenty police detectives working in one of the largest Australian police organizations. It suggests that police detectives in this study have negative perceptions of diversity, and associate diversity with unfair advantages in the workplace. In Australian culture, the phrase ‘rorting the system’ is an informal expression used to describe individuals or groups of people who take unfair advantage of a public service or workplace policy to change their circumstances. The findings suggest that detectives in this study believe diversity enables some officers to take advantage of workplace policy and ‘rort’ the system.

Transboundary water cooperation and conflict resolution in the Southern African region: influence of the 1890 Anglo-Germany Treaty

The availability, distribution and control of freshwater resources have been at the centre of the human story since the start of the Neolithic revolution roughly 12,000 years ago. With the advent of the modern nation state and its attendant emphasis on sovereignty, self-sufficiency and rivalry, it comes as no surprise that interactions between states over shared watercourses have at times been tense and conflictual. This fact was elaborated by the Ex- UN Secretary General; Kofi Annan, Message during the World Water Day on 22nd March, 2002. He warned that… “Fierce national competition over water resources has prompted fears that water issues contain the seeds of violent conflict. By the year 2025 two thirds of the world’s population is likely to live in countries with moderate or severe water shortages as demand for water approaches the limit of the available supply”. Water as a fugitive resource, respects neither political boundaries nor commonly accepted notions of fairness or equity, hence posed the most complex management challenges to water managers of today. In the SADC region, shared waters cannot be viewed in a purely national context due to its fluidity and the mobility of its nature. It is factual that, over 70% of the water bodies in the region are transboundary in nature. In terms of state practice, the concept of community of interest is commonly traced back to a French decree of 1792 dealing with the opening of the Scheldt River to Navigation. The position expressed in this decree was quickly adopted in a number of instruments concerned primarily with rights of navigation in international rivers, but also in some early agreements not restricted to navigational uses. Therefore, the lakes, and watercourses which form the frontier between the two states or which are situated at the territory of both or which flow into the said lakes and watercourses shall continue to be considered as “common’. In this regard one may wish to refer to the recent global instruments namely; the UN Convention on the Law of the Non-Navigational uses of International Water (1997) which came into force on 17th August 2014 and the Convention on the Protection and Uses of Transboundary Watercourses and International Lakes (1992) which came into force on 6th October, 1996 and further in 2016 became an official global legal framework for transboundary water cooperation. These instruments are regarded as a vital step in building a strong foundation for global principles on water management and governance. Legal agreements between states during the colonial era as well as post-independence in the Southern Africa region, have formed the bedrock of cooperative water resources management regionally. The Anglo Germany Treaty of July, 1890 (The Helgoland Treaty), had established an agreement between the colonial powers of Great Britain, France, Portugal, Belgium and Germany and their respective spheres of influence over the African nations aimed to establish borders between the nations. Interesting to note in the presence of scarcity of geo-information over the areas in question; the water bodies (Rivers and Lakes) were used to mark the lines of influence hence boundaries of the sovereign states of today. This chapter therefore, will provide an account of the influence of the 1890 Anglo – Germany Treaty (Helgoland Treaty) and international customary law in regard to conflict resolution and transboundary water cooperation in the Southern Africa Region (SADC). It will also examine some of available information as well as the historical background of boundary treaties; legal frameworks for cooperation; importance of Africa Union(AU) resolutions on the same, such as Resolution AHG/Res16(1) of July 1964 as well as resolution CM/Res.1069(XLIV) of 1986 and finally a conclusion.

Fair active learning

Machine learning (ML) is increasingly being used in high-stakes applications impacting society. Therefore, it is of critical importance that ML models do not propagate discrimination. Collecting accurate labeled data in societal applications is challenging and costly. Active learning is a promising approach to build an accurate classifier by interactively querying an oracle within a labeling budget. We introduce the fair active learning framework to carefully select data points to be labeled so as to balance model accuracy and fairness. To incorporate the notion of fairness in the active learning sampling core, it is required to measure the fairness of the model after adding each unlabeled sample. Since their labels are unknown in advance, we propose an expected fairness metric to probabilistically measure the impact of each sample if added for each possible class label. Next, we propose multiple optimizations to balance the trade-off between accuracy and fairness. Our first optimization linearly aggregate the expected fairness with entropy using a control parameter. To avoid erroneous estimation of the expected fairness, we propose a nested approach to maintain the accuracy of the model, limiting the search space to the top bucket of sample points with large entropy. Finally, to ensure the unfairness reduction of the model after labeling, we propose to replicate the points that truly reduce the unfairness after labeling. We demonstrate the effectiveness and efficiency of our proposed algorithms over widely used benchmark datasets using demographic parity and equalized odds notions of fairness.

Some Philosophical Questions to Understand the Role of Arbitrators through the Notion of Fair and Equitable Treatment

This Article explores the contributions that philosophical debates can make to understanding the role of arbitrators and international adjudicators in investment disciplines. To do so, the article surveys a series of decisions and discussions in the doctrine and arbitration award concerning the standard of fair an equitable treatment.

Agree to Disagree: Subjective Fairness in Privacy-Restricted Decentralised Conflict Resolution.

Fairness is commonly seen as a property of the global outcome of a system and assumes centralisation and complete knowledge. However, in real decentralised applications, agents only have partial observation capabilities. Under limited information, agents rely on communication to divulge some of their private (and unobservable) information to others. When an agent deliberates to resolve conflicts, limited knowledge may cause its perspective of a correct outcome to differ from the actual outcome of the conflict resolution. This is subjective unfairness. As human systems and societies are organised by rules and norms, hybrid human-agent and multi-agent environments of the future will require agents to resolve conflicts in a decentralised and rule-aware way. Prior work achieves such decentralised, rule-aware conflict resolution through cultures: explainable architectures that embed human regulations and norms via argumentation frameworks with verification mechanisms. However, this prior work requires agents to have full state knowledge of each other, whereas many distributed applications in practice admit partial observation capabilities, which may require agents to communicate and carefully opt to release information if privacy constraints apply. To enable decentralised, fairness-aware conflict resolution under privacy constraints, we have two contributions: 1) a novel interaction approach and 2) a formalism of the relationship between privacy and fairness. Our proposed interaction approach is an architecture for privacy-aware explainable conflict resolution where agents engage in a dialogue of hypotheses and facts. To measure the privacy-fairness relationship, we define subjective and objective fairness on both the local and global scope and formalise the impact of partial observability due to privacy in these different notions of fairness. We first study our proposed architecture and the privacy-fairness relationship in the abstract, testing different argumentation strategies on a large number of randomised cultures. We empirically demonstrate the trade-off between privacy, objective fairness, and subjective fairness and show that better strategies can mitigate the effects of privacy in distributed systems. In addition to this analysis across a broad set of randomised abstract cultures, we analyse a case study for a specific scenario: we instantiate our architecture in a multi-agent simulation of prioritised rule-aware collision avoidance with limited information disclosure.

Achieving energy justice in Malawi: from key challenges to policy recommendations

Addressing energy provision and access in Sub-Saharan Africa is a key global challenge. Drawing on interviews with key stakeholders, this paper applies an energy justice framework in overviewing energy realities and policies in Malawi, where electricity access remains among the lowest in Sub-Saharan Africa. The use of woodfuel remains high for meeting cooking, heating, and lighting needs leading to indoor air pollution, with serious health consequences, and widespread deforestation. Responses to these dual challenges, a lack of electricity access and ongoing woodfuel use, must be rooted in notions of equity, fairness, and justice. Application of energy justice theorising provides insights into how policy stakeholders are responding to complex and interconnected issues of energy generation and access in low-income settings. Overall, a just response to these energy challenges is possible, but only if it is built on local inclusive governance with fairer and effective systems of investment.

Algorithmic Redistricting and Black Representation in US Elections

In the United States, the careful crafting of electoral districts has been a powerful tool for politicians to limit groups’ political power or exclude them from representation entirely, most prominently to the detriment of political parties and racial and ethnic minority communities. Beginning in the 1960s, experts began proposing algorithmic solutions to the redistricting problem, in which a “neutral” computer program could draw “fair” districts free of human influence. Despite the traction that these proposals achieved both in academic and popular discourse, little work has been done to understand the extent to which algorithmically drawn districts do or do not comport with notions of fairness and equity. In this work, we perform such an analysis, running several proposed algorithms to generate districts in Alabama and Michigan. We observe that in both of these states, all four algorithms generate plans that provide fewer districts where Black voters would be expected to decide the outcome of the election, relative to both the proportion of Black people in their populaces as well as to the number of Black opportunity districts in the plans actually enacted by their state legislatures. We conclude with some discussion about the role of algorithms in redistricting moving forward, and how these tools might be used to enhance, rather than restrict, the ability for various communities to achieve elected representation.

Dropping the GRE, keeping the GRE, or GRE-optional admissions? Considering tradeoffs and fairness

In considering whether to retain the GRE in graduate school admissions, admissions committees often pursue two objectives: (a) performance in graduate school (e.g., admitting individuals who will perform better in classes and research), and (b) diversity/fairness (e.g., equal selection rates between demographic groups). Drawing upon HR research (adverse impact research), we address four issues in using the GRE. First, we review the tension created between two robust findings: (a) validity of the GRE for predicting graduate school performance (rooted in the principle of standardization and a half-century of educational and psychometric research), and (b) the achievement gap in test scores between demographic groups (rooted in several centuries of systemic racism). This empirical tension can often produce a local diversity-performance tradeoff for admissions committees. Second, we use Pareto-optimal tradeoff curves to formalize potential diversity-performance tradeoffs, guiding how much weight to assign the GRE in admissions. Whether dropping the GRE produces suboptimal admissions depends upon one’s relative valuation of diversity versus performance. Third, we review three distinct notions of test fairness—equality, test equity, and performance equity—which have differing implications for dropping the GRE. Finally, we consider test fairness under GRE-optional admissions, noting the missing data problem when GRE is an incomplete variable. Supplemental data for this article is available online at

Negligence is not ignorance

ABSTRACT Recent interest in the epistemic condition on moral responsibility has raised a new challenge to the view that persons are directly responsible for negligent conduct. According to an influential argument, the epistemic condition on responsibility requires genuine, clear eyed akrasia. All other kinds of wrongdoing, including negligence, constitute ignorant wrongdoing. My aim in this paper is to defend direct responsibility for negligence against this challenge by arguing that negligence does not constitute ignorant wrongdoing. In distinguishing negligence from ignorant wrongdoing, I do not dispute the idea that negligence is involves unawareness. Rather, I argue that there is a difference between failing to bring a relevant fact to mind at the appropriate time and being genuinely ignorant of some relevant fact. Standard cases of negligence involve failing to bring a relevant fact to mind, not genuine ignorance. To support the significance of this distinction, I appeal to the notion of fairness. Fairness supports understanding the epistemic condition on responsibility as requiring not occurrent beliefs but rather reasonable internal access to the relevant beliefs and thereby enables me to defend direct responsibility for negligence.

Online exam proctoring technologies: Educational innovation or deterioration?

AbstractDuring the coronavirus disease‐2019 (Covid‐19) pandemic, many universities have adopted online exam proctoring technologies to monitor and control an increasing number of student cheating incidents. Although it looks like a natural and effective solution for a fair assessment of student online learning performance, the authors argue that proctoring technologies are rooted in problematic assumptions about educational fairness and authoritarian pedagogical approaches. The authors have conducted a qualitative case study in a large‐sized, top‐tier university in South Korea to investigate the negative impacts of adopting proctoring technologies on student subjectivities, pedagogical relationships and educational outcomes, which have not been fully discussed in previous studies. By utilising Foucault's theorisation of disciplinary governmentality, the authors effectively demonstrate that the binary subjectification of students as cheaters and the cheated has degraded the value of student engagement in university education whilst creating more competitive and distrusting relationships amongst students and between students and teachers. Nevertheless, without challenging the unethical consequences of online proctoring technologies or fundamentally unfair social and educational systems, students willingly accept and adopt them as docile bodies, which has led to educational deterioration rather than innovation. Practitioner notesWhat is already known about this topic There is an increasing number of online exam proctoring technologies available with advanced technical features. During the coronavirus disease‐2019 (Covid‐19) pandemic, many universities have adopted online exam proctoring technologies with exam‐related policies and regulations to stop student cheating behaviours. Previous studies have discussed both advantages and disadvantages of adopting online exam proctoring technologies, including specific ethical concerns. What this paper adds Online exam proctoring technologies are deeply rooted in problematic educational approaches such as teacher‐centred knowledge transmission. The adoption of online exam proctoring technologies has produced negative impacts on student subjectivities, pedagogical relationships and educational outcomes. Focusing on student cheating as an individual and interpersonal problem neglects the more fundamental issue of social and educational inequality. Implications for practice and/or policy The negative consequences and damages created by the adoption of online exams and online exam proctoring technologies need to be carefully reflected. The notion of academic fairness needs to be approached with a broader perspective, considering the different social and academic circumstances each student is in. It is essential to critically engage students with the broader conversation about educational fairness to develop them as critical thinkers and future leaders in their chosen fields.

Reassessing privacy, fairness and governance in the age of algorithms and the impact on society and institutions

While algorithms will play a central role in the future of business and society, there is a lack of awareness of the grave implications to fundamental human values and to institutions that are built around these values. Algorithms have already rendered the present framework of law and governance outdated. While new technologies have historically roiled societies with the privacy conundrum at times of great change, the ability to gather, analyse, and combine vast quantities of data from different sources using powerful algorithms, impose a significant challenge to the traditional notion of privacy and fairness, as never seen before. This paper elucidates the concepts of data and algorithms along with our current notions of privacy and fairness. It explores the multi-dimensional algorithmic challenges to our present notion of privacy and fairness and to institutions to which privacy and fairness are at their core. It also deliberates on safeguards and remedial actions being considered at present.

Fair Contrastive Learning on Graphs

Node representation learning plays a critical role in learning over graphs. Specifically, the success of contrastive learning methods in unsupervised node representation learning has been demonstrated for various tasks, which has led to increase in attention towards the field. Despite the increasing popularity, fairness is widely under-explored in the area. Motivated by this, this study proposes novel fairness-aware graph augmentations based on adaptive feature masking and edge deletion, in order to mitigate the bias in graph contrastive learning. Different fairness notions on graphs are introduced in the study to guide the designs of the proposed adaptive augmentation schemes. Moreover, it is quantitatively shown that the proposed feature masking scheme can reduce the intrinsic bias. Experimental results on four real-world networks are presented to show that the introduced augmentation frameworks can improve group fairness measures together with comparable classification accuracy to state-of-the-art graph contrastive learning studies for node classification.

Social impacts of algorithmic decision-making: A research agenda for the social sciences

Academic and public debates are increasingly concerned with the question whether and how algorithmic decision-making (ADM) may reinforce social inequality. Most previous research on this topic originates from computer science. The social sciences, however, have huge potentials to contribute to research on social consequences of ADM. Based on a process model of ADM systems, we demonstrate how social sciences may advance the literature on the impacts of ADM on social inequality by uncovering and mitigating biases in training data, by understanding data processing and analysis, as well as by studying social contexts of algorithms in practice. Furthermore, we show that fairness notions need to be evaluated with respect to specific outcomes of ADM systems and with respect to concrete social contexts. Social sciences may evaluate how individuals handle algorithmic decisions in practice and how single decisions aggregate to macro social outcomes. In this overview, we highlight how social sciences can apply their knowledge on social stratification and on substantive domains of ADM applications to advance the understanding of social impacts of ADM.

Fairness and Cost Constrained Privacy-Aware Record Linkage

Record linkage algorithms match and link records from different databases that refer to the same real-world entity based on direct and/or quasi-identifiers, such as name, address, age, and gender, available in the records. Since these identifiers generally contain personal identifiable information (PII) about the entities, record linkage algorithms need to be developed with privacy constraints. Known as privacy-preserving record linkage (PPRL), many research studies have been conducted to perform the linkage on encoded and/or encrypted identifiers. Differential privacy (DP) combined with computationally efficient encoding methods, e.g. Bloom filter encoding, has been used to develop PPRL with provable privacy guarantees. The standard DP notion does not however address other constraints, among which the most important ones are fairness-bias and cost of linkage in terms of number of record pairs to be compared. In this work, we propose new notions of fairness-constrained DP and fairness and cost-constrained DP for PPRL and develop a framework for PPRL with these new notions of DP combined with Bloom filter encoding. We provide theoretical proofs for the new DP notions for fairness and cost-constrained PPRL and experimentally evaluate them on two datasets containing person-specific data. Our experimental results show that with these new notions of DP, PPRL with better performance (compared to the standard DP notion for PPRL) can be achieved with regard to privacy, cost and fairness constraints.

A review of causality-based fairness machine learning

With the wide application of machine learning driven automated decisions (e.g., education, loan approval, and hiring) in daily life, it is critical to address the problem of discriminatory behavior toward certain individuals or groups. Early studies focused on defining the correlation/association-based notions, such as statistical parity, equalized odds, etc. However, recent studies reflect that it is necessary to use causality to address the problem of fairness. This review provides an exhaustive overview of notions and methods for detecting and eliminating algorithmic discrimination from a causality perspective. The review begins by introducing the common causality-based definitions and measures for fairness. We then review causality-based fairness-enhancing methods from the perspective of pre-processing, in-processing and post-processing mechanisms, and conduct a comprehensive analysis of the advantages, disadvantages, and applicability of these mechanisms. In addition, this review also examines other domains where researchers have observed unfair outcomes and the ways they have tried to address them. There are still many challenges that hinder the practical application of causality-based fairness notions, specifically the difficulty of acquiring causal graphs and identifiability of causal effects. One of the main purposes of this review is to spark more researchers to tackle these challenges in the near future.