arXiv subjects

Julia Stoyanovich

Publications and source records attributed to Julia Stoyanovich.

At least 37 records · Page 2Linked to original sources

arXiv2025

CREDAL: Close Reading of Data Models

Data models are necessary for the birth of data and of any data-driven system. Indeed, every algorithm, every machine learning model, every statistical model, and every database has an underlying data model without which the system would not be usable. Hence, data models are excellent sites for interrogating the (material, social, political, ...) conditions giving rise to a data system. Towards this, drawing inspiration from literary criticism, we propose to closely read data models in the same spirit as we closely read literary artifacts. Close readings of data models reconnect us with, among other things, the materiality, the genealogies, the techne, the closed nature, and the design of technical systems. While recognizing from literary theory that there is no one correct way to read, it is nonetheless critical to have systematic guidance for those unfamiliar with close readings. This is especially true for those trained in the computing and data sciences, who too often are enculturated to set aside the socio-political aspects of data work. A systematic methodology for reading data models currently does not exist. To fill this gap, we present the CREDAL methodology for close readings of data models. We detail our iterative development process and present results of a qualitative evaluation of CREDAL demonstrating its usability, usefulness, and effectiveness in the critical study of data.

cs.DB↗

arXiv2025

Safeguarding Large Language Models in Real-time with Tunable Safety-Performance Trade-offs

Large Language Models (LLMs) have been shown to be susceptible to jailbreak attacks, or adversarial attacks used to illicit high risk behavior from a model. Jailbreaks have been exploited by cybercriminals and blackhat actors to cause significant harm, highlighting the critical need to safeguard widely-deployed models. Safeguarding approaches, which include fine-tuning models or having LLMs "self-reflect", may lengthen the inference time of a model, incur a computational penalty, reduce the semantic fluency of an output, and restrict ``normal'' model behavior. Importantly, these Safety-Performance Trade-offs (SPTs) remain an understudied area. In this work, we introduce a novel safeguard, called SafeNudge, that combines Controlled Text Generation with "nudging", or using text interventions to change the behavior of a model. SafeNudge triggers during text-generation while a jailbreak attack is being executed, and can reduce successful jailbreak attempts by 30% by guiding the LLM towards a safe responses. It adds minimal latency to inference and has a negligible impact on the semantic fluency of outputs. Further, we allow for tunable SPTs. SafeNudge is open-source and available through https://pypi.org/, and is compatible with models loaded with the Hugging Face "transformers" library.

cs.CL↗

arXiv2024

Using Case Studies to Teach Responsible AI to Industry Practitioners

Responsible AI (RAI) encompasses the science and practice of ensuring that AI design, development, and use are socially sustainable -- maximizing the benefits of technology while mitigating its risks. Industry practitioners play a crucial role in achieving the objectives of RAI, yet there is a persistent a shortage of consolidated educational resources and effective methods for teaching RAI to practitioners. In this paper, we present a stakeholder-first educational approach using interactive case studies to foster organizational and practitioner-level engagement and enhance learning about RAI. We detail our partnership with Meta, a global technology company, to co-develop and deliver RAI workshops to a diverse company audience. Assessment results show that participants found the workshops engaging and reported an improved understanding of RAI principles, along with increased motivation to apply them in their work.

cs.HC↗

arXiv2024

Making Transparency Advocates: An Educational Approach Towards Better Algorithmic Transparency in Practice

Concerns about the risks and harms posed by artificial intelligence (AI) have resulted in significant study into algorithmic transparency, giving rise to a sub-field known as Explainable AI (XAI). Unfortunately, despite a decade of development in XAI, an existential challenge remains: progress in research has not been fully translated into the actual implementation of algorithmic transparency by organizations. In this work, we test an approach for addressing the challenge by creating transparency advocates, or motivated individuals within organizations who drive a ground-up cultural shift towards improved algorithmic transparency. Over several years, we created an open-source educational workshop on algorithmic transparency and advocacy. We delivered the workshop to professionals across two separate domains to improve their algorithmic transparency literacy and willingness to advocate for change. In the weeks following the workshop, participants applied what they learned, such as speaking up for algorithmic transparency at an organization-wide AI strategy meeting. We also make two broader observations: first, advocacy is not a monolith and can be broken down into different levels. Second, individuals' willingness for advocacy is affected by their professional field. For example, news and media professionals may be more likely to advocate for algorithmic transparency than those working at technology start-ups.

cs.CY↗

arXiv2024

Are Data Experts Buying into Differentially Private Synthetic Data? Gathering Community Perspectives

Data privacy is a core tenet of responsible computing, and in the United States, differential privacy (DP) is the dominant technical operationalization of privacy-preserving data analysis. With this study, we qualitatively examine one class of DP mechanisms: private data synthesizers. To that end, we conducted semi-structured interviews with data experts: academics and practitioners who regularly work with data. Broadly, our findings suggest that quantitative DP benchmarks must be grounded in practitioner needs, while communication challenges persist. Participants expressed a need for context-aware DP solutions, focusing on parity between research outcomes on real and synthetic data. Our analysis led to three recommendations: (1) improve existing insufficient sanitized benchmarks; successful DP implementations require well-documented, partner-vetted use cases, (2) organizations using DP synthetic data should publish discipline-specific standards of evidence, and (3) tiered data access models could allow researchers to gradually access sensitive data based on demonstrated competence with high-privacy, low-fidelity synthetic data.

cs.HC↗

arXiv2024

Query Refinement for Diverse Top-$k$ Selection

Database queries are often used to select and rank items as decision support for many applications. As automated decision-making tools become more prevalent, there is a growing recognition of the need to diversify their outcomes. In this paper, we define and study the problem of modifying the selection conditions of an ORDER BY query so that the result of the modified query closely fits some user-defined notion of diversity while simultaneously maintaining the intent of the original query. We show the hardness of this problem and propose a Mixed Integer Linear Programming (MILP) based solution. We further present optimizations designed to enhance the scalability and applicability of the solution in real-life scenarios. We investigate the performance characteristics of our algorithm and show its efficiency and the usefulness of our optimizations.

cs.DB↗

arXiv2024

Fairness in Algorithmic Recourse Through the Lens of Substantive Equality of Opportunity

Algorithmic recourse -- providing recommendations to those affected negatively by the outcome of an algorithmic system on how they can take action and change that outcome -- has gained attention as a means of giving persons agency in their interactions with artificial intelligence (AI) systems. Recent work has shown that even if an AI decision-making classifier is ``fair'' (according to some reasonable criteria), recourse itself may be unfair due to differences in the initial circumstances of individuals, compounding disparities for marginalized populations and requiring them to exert more effort than others. There is a need to define more methods and metrics for evaluating fairness in recourse that span a range of normative views of the world, and specifically those that take into account time. Time is a critical element in recourse because the longer it takes an individual to act, the more the setting may change due to model or data drift. This paper seeks to close this research gap by proposing two notions of fairness in recourse that are in normative alignment with substantive equality of opportunity, and that consider time. The first considers the (often repeated) effort individuals exert per successful recourse event, and the second considers time per successful recourse event. Building upon an agent-based framework for simulating recourse, this paper demonstrates how much effort is needed to overcome disparities in initial circumstances. We then proposes an intervention to improve the fairness of recourse by rewarding effort, and compare it to existing strategies.

cs.LG↗

arXiv2024

A New Paradigm for Counterfactual Reasoning in Fairness and Recourse

Counterfactuals and counterfactual reasoning underpin numerous techniques for auditing and understanding artificial intelligence (AI) systems. The traditional paradigm for counterfactual reasoning in this literature is the interventional counterfactual, where hypothetical interventions are imagined and simulated. For this reason, the starting point for causal reasoning about legal protections and demographic data in AI is an imagined intervention on a legally-protected characteristic, such as ethnicity, race, gender, disability, age, etc. We ask, for example, what would have happened had your race been different? An inherent limitation of this paradigm is that some demographic interventions -- like interventions on race -- may not translate into the formalisms of interventional counterfactuals. In this work, we explore a new paradigm based instead on the backtracking counterfactual, where rather than imagine hypothetical interventions on legally-protected characteristics, we imagine alternate initial conditions while holding these characteristics fixed. We ask instead, what would explain a counterfactual outcome for you as you actually are or could be? This alternate framework allows us to address many of the same social concerns, but to do so while asking fundamentally different questions that do not rely on demographic interventions.

cs.AI↗

arXiv2023

A Simple and Practical Method for Reducing the Disparate Impact of Differential Privacy

Differentially private (DP) mechanisms have been deployed in a variety of high-impact social settings (perhaps most notably by the U.S. Census). Since all DP mechanisms involve adding noise to results of statistical queries, they are expected to impact our ability to accurately analyze and learn from data, in effect trading off privacy with utility. Alarmingly, the impact of DP on utility can vary significantly among different sub-populations. A simple way to reduce this disparity is with stratification. First compute an independent private estimate for each group in the data set (which may be the intersection of several protected classes), then, to compute estimates of global statistics, appropriately recombine these group estimates. Our main observation is that naive stratification often yields high-accuracy estimates of population-level statistics, without the need for additional privacy budget. We support this observation theoretically and empirically. Our theoretical results center on the private mean estimation problem, while our empirical results center on extensive experiments on private data synthesis to demonstrate the effectiveness of stratification on a variety of private mechanisms. Overall, we argue that this straightforward approach provides a strong baseline against which future work on reducing utility disparities of DP mechanisms should be compared.

cs.CR↗

arXiv2023

Setting the Right Expectations: Algorithmic Recourse Over Time

Algorithmic systems are often called upon to assist in high-stakes decision making. In light of this, algorithmic recourse, the principle wherein individuals should be able to take action against an undesirable outcome made by an algorithmic system, is receiving growing attention. The bulk of the literature on algorithmic recourse to-date focuses primarily on how to provide recourse to a single individual, overlooking a critical element: the effects of a continuously changing context. Disregarding these effects on recourse is a significant oversight, since, in almost all cases, recourse consists of an individual making a first, unfavorable attempt, and then being given an opportunity to make one or several attempts at a later date - when the context might have changed. This can create false expectations, as initial recourse recommendations may become less reliable over time due to model drift and competition for access to the favorable outcome between individuals. In this work we propose an agent-based simulation framework for studying the effects of a continuously changing environment on algorithmic recourse. In particular, we identify two main effects that can alter the reliability of recourse for individuals represented by the agents: (1) competition with other agents acting upon recourse, and (2) competition with new agents entering the environment. Our findings highlight that only a small set of specific parameterizations result in algorithmic recourse that is reliable for agents over time. Consequently, we argue that substantial additional work is needed to understand recourse reliability over time, and to develop recourse methods that reward agents' effort.

cs.LG↗

arXiv2023

Epistemic Parity: Reproducibility as an Evaluation Metric for Differential Privacy

Differential privacy (DP) data synthesizers support public release of sensitive information, offering theoretical guarantees for privacy but limited evidence of utility in practical settings. Utility is typically measured as the error on representative proxy tasks, such as descriptive statistics, accuracy of trained classifiers, or performance over a query workload. The ability for these results to generalize to practitioners' experience has been questioned in a number of settings, including the U.S. Census. In this paper, we propose an evaluation methodology for synthetic data that avoids assumptions about the representativeness of proxy tasks, instead measuring the likelihood that published conclusions would change had the authors used synthetic data, a condition we call epistemic parity. Our methodology consists of reproducing empirical conclusions of peer-reviewed papers on real, publicly available data, then re-running these experiments a second time on DP synthetic data, and comparing the results. We instantiate our methodology over a benchmark of recent peer-reviewed papers that analyze public datasets in the ICPSR repository. We model quantitative claims computationally to automate the experimental workflow, and model qualitative claims by reproducing visualizations and comparing the results manually. We then generate DP synthetic datasets using multiple state-of-the-art mechanisms, and estimate the likelihood that these conclusions will hold. We find that state-of-the-art DP synthesizers are able to achieve high epistemic parity for several papers in our benchmark. However, some papers, and particularly some specific findings, are difficult to reproduce for any of the synthesizers. We advocate for a new class of mechanisms that favor stronger utility guarantees and offer privacy protection with a focus on application-specific threat models and risk-assessment.

cs.CR↗

arXiv2023

Most Expected Winner: An Interpretation of Winners over Uncertain Voter Preferences

It remains an open question how to determine the winner of an election when voter preferences are incomplete or uncertain. One option is to assume some probability space over the voting profile and select the Most Probable Winner (MPW) -- the candidate or candidates with the best chance of winning. In this paper, we propose an alternative winner interpretation, selecting the Most Expected Winner (MEW) according to the expected performance of the candidates. We separate the uncertainty in voter preferences into the generation step and the observation step, which gives rise to a unified voting profile combining both incomplete and probabilistic voting profiles. We use this framework to establish the theoretical hardness of \mew over incomplete voter preferences, and then identify a collection of tractable cases for a variety of voting profiles, including those based on the popular Repeated Insertion Model (RIM) and its special case, the Mallows model. We develop solvers customized for various voter preference types to quantify the candidate performance for the individual voters, and propose a pruning strategy that optimizes computation. The performance of the proposed solvers and pruning strategy is evaluated extensively on real and synthetic benchmarks, showing that our methods are practical.

cs.GT↗

arXiv2023

The Unbearable Weight of Massive Privilege: Revisiting Bias-Variance Trade-Offs in the Context of Fair Prediction

In this paper we revisit the bias-variance decomposition of model error from the perspective of designing a fair classifier: we are motivated by the widely held socio-technical belief that noise variance in large datasets in social domains tracks demographic characteristics such as gender, race, disability, etc. We propose a conditional-iid (ciid) model built from group-specific classifiers that seeks to improve on the trade-offs made by a single model (iid setting). We theoretically analyze the bias-variance decomposition of different models in the Gaussian Mixture Model, and then empirically test our setup on the COMPAS and folktables datasets. We instantiate the ciid model with two procedures that improve "fairness" by conditioning out undesirable effects: first, by conditioning directly on sensitive attributes, and second, by clustering samples into groups and conditioning on cluster membership (blind to protected group membership). Our analysis suggests that there might be principled procedures and concrete real-world use cases under which conditional models are preferred, and our striking empirical results strongly indicate that non-iid settings, such as the ciid setting proposed here, might be more suitable for big data applications in social contexts.

cs.LG↗

arXiv2023

The Possibility of Fairness: Revisiting the Impossibility Theorem in Practice

The ``impossibility theorem'' -- which is considered foundational in algorithmic fairness literature -- asserts that there must be trade-offs between common notions of fairness and performance when fitting statistical models, except in two special cases: when the prevalence of the outcome being predicted is equal across groups, or when a perfectly accurate predictor is used. However, theory does not always translate to practice. In this work, we challenge the implications of the impossibility theorem in practical settings. First, we show analytically that, by slightly relaxing the impossibility theorem (to accommodate a \textit{practitioner's} perspective of fairness), it becomes possible to identify a large set of models that satisfy seemingly incompatible fairness constraints. Second, we demonstrate the existence of these models through extensive experiments on five real-world datasets. We conclude by offering tools and guidance for practitioners to understand when -- and to what degree -- fairness along multiple criteria can be achieved. For example, if one allows only a small margin-of-error between metrics, there exists a large set of models simultaneously satisfying \emph{False Negative Rate Parity}, \emph{False Positive Rate Parity}, and \emph{Positive Predictive Value Parity}, even when there is a moderate prevalence difference between groups. This work has an important implication for the community: achieving fairness along multiple metrics for multiple groups (and their intersections) is much more possible than was previously believed.

cs.LG↗

arXiv2022

Counterfactuals for the Future

Counterfactuals are often described as 'retrospective,' focusing on hypothetical alternatives to a realized past. This description relates to an often implicit assumption about the structure and stability of exogenous variables in the system being modeled -- an assumption that is reasonable in many settings where counterfactuals are used. In this work, we consider cases where we might reasonably make a different assumption about exogenous variables, namely, that the exogenous noise terms of each unit do exhibit some unit-specific structure and/or stability. This leads us to a different use of counterfactuals -- a 'forward-looking' rather than 'retrospective' counterfactual. We introduce "counterfactual treatment choice," a type of treatment choice problem that motivates using forward-looking counterfactuals. We then explore how mismatches between interventional versus forward-looking counterfactual approaches to treatment choice, consistent with different assumptions about exogenous noise, can lead to counterintuitive results.

cs.AI↗

arXiv2022

Disaggregated Interventions to Reduce Inequality

A significant body of research in the data sciences considers unfair discrimination against social categories such as race or gender that could occur or be amplified as a result of algorithmic decisions. Simultaneously, real-world disparities continue to exist, even before algorithmic decisions are made. In this work, we draw on insights from the social sciences brought into the realm of causal modeling and constrained optimization, and develop a novel algorithmic framework for tackling pre-existing real-world disparities. The purpose of our framework, which we call the "impact remediation framework," is to measure real-world disparities and discover the optimal intervention policies that could help improve equity or access to opportunity for those who are underserved with respect to an outcome of interest. We develop a disaggregated approach to tackling pre-existing disparities that relaxes the typical set of assumptions required for the use of social categories in structural causal models. Our approach flexibly incorporates counterfactuals and is compatible with various ontological assumptions about the nature of social categories. We demonstrate impact remediation with a hypothetical case study and compare our disaggregated approach to an existing state-of-the-art approach, comparing its structure and resulting policy recommendations. In contrast to most work on optimal policy learning, we explore disparity reduction itself as an objective, explicitly focusing the power of algorithms on reducing inequality.

cs.LG↗

arXiv2022

Rankers, Rankees, & Rankings: Peeking into the Pandora's Box from a Socio-Technical Perspective

Algorithmic rankers have a profound impact on our increasingly data-driven society. From leisurely activities like the movies that we watch, the restaurants that we patronize; to highly consequential decisions, like making educational and occupational choices or getting hired by companies -- these are all driven by sophisticated yet mostly inaccessible rankers. A small change to how these algorithms process the rankees (i.e., the data items that are ranked) can have profound consequences. For example, a change in rankings can lead to deterioration of the prestige of a university or have drastic consequences on a job candidate who missed out being in the list of the preferred top-k for an organization. This paper is a call to action to the human-centered data science research community to develop principled methods, measures, and metrics for studying the interactions among the socio-technical context of use, technological innovations, and the resulting consequences of algorithmic rankings on multiple stakeholders. Given the spate of new legislations on algorithmic accountability, it is imperative that researchers from social science, human-computer interaction, and data science work in unison for demystifying how rankings are produced, who has agency to change them, and what metrics of socio-technical impact one must use for informing the context of use.

cs.HC↗

arXiv2022

Fairness in Ranking: A Survey

In the past few years, there has been much work on incorporating fairness requirements into algorithmic rankers, with contributions coming from the data management, algorithms, information retrieval, and recommender systems communities. In this survey we give a systematic overview of this work, offering a broad perspective that connects formalizations and algorithmic approaches across subfields. An important contribution of our work is in developing a common narrative around the value frameworks that motivate specific fairness-enhancing interventions in ranking. This allows us to unify the presentation of mitigation objectives and of algorithmic techniques to help meet those objectives or identify trade-offs. In this survey, we describe four classification frameworks for fairness-enhancing interventions, along which we relate the technical methods surveyed in this paper, discuss evaluation datasets, and present technical work on fairness in score-based ranking. Then, we present methods that incorporate fairness in supervised learning, and also give representative examples of recent work on fairness in recommendation and matchmaking systems. We also discuss evaluation frameworks for fair score-based ranking and fair learning-to-rank, and draw a set of recommendations for the evaluation of fair ranking methods.

cs.IR↗