Recent AI trends seek to align AI models to learned human-centric objectives, such as personal preferences, utility, or societal values. Using standard preference elicitation methods, researchers and practitioners build models of human decisions and judgments, to which AI models are aligned. However, standard elicitation methods often fail to capture the true cognitive processes behind human decision making, such as the use of heuristics or simplifying structured thought patterns. To address this limitation, we take an axiomatic approach to learning cognitively faithful decision processes from pairwise comparisons. Building on the vast literature characterizing cognitive processes that contribute to human decision-making and pairwise comparisons, we derive a class of models in which individual features are first processed with learned rules, then aggregated via a fixed rule, such as the Bradley-Terry rule, to produce a decision. This structured processing of information ensures that such models are realistic and feasible candidates to represent underlying human decision-making processes. We demonstrate the efficacy of this modeling approach by learning interpretable models of human decision making in a kidney allocation task, and show that our proposed models match or surpass the accuracy of prior models of human pairwise decision-making.

Alignment methods in moral domains seek to elicit moral preferences of human stakeholders and incorporate them into AI. This presupposes moral preferences as static targets, but such preferences often evolve over time. Proper alignment of AI to dynamic human preferences should ideally account for "legitimate" changes to moral reasoning, while ignoring changes related to attention deficits, cognitive biases, or other arbitrary factors. However, common AI alignment approaches largely neglect temporal changes in preferences, posing serious challenges to proper alignment, especially in high-stakes applications of AI, e.g., in healthcare domains, where misalignment can jeopardize the trustworthiness of the system and yield serious individual and societal harms. This work investigates the extent to which people's moral preferences change over time, and the impact of such changes on AI alignment. Our study is grounded in the kidney allocation domain, where we elicit responses to pairwise comparisons of hypothetical kidney transplant patients from over 400 participants across 3-5 sessions. We find that, on average, participants change their response to the same scenario presented at different times around 6-20% of the time (exhibiting "response instability"). Additionally, we observe significant shifts in several participants' retrofitted decision-making models over time (capturing "model instability"). Predictive performance of simple AI models decreases as a function of both response and model instability. Moreover, predictive performance diminishes over time, highlighting the importance of accounting for temporal changes in preferences during training. These findings raise fundamental normative and technical challenges relevant to AI alignment, highlighting the need to better understand the object of alignment (what to align to) when user preferences change significantly over time, including the mechanisms underlying these changes.

AI systems are increasingly employed as decision aids, decision delegates, or autonomous decision-makers. This position paper argues that in many settings, particularly high-stakes decision-making, we need accurate cognitively-aligned AI systems that reason similarly to their users, and faithfully communicate their reasoning. We review evidence that cognitive alignment improves understandability and trustworthiness, and provide new survey data showing that many users find cognitive alignment “essential” when an AI’s rationale for a judgment or action is important to them. We outline the gaps between existing alignment methods and what is needed to achieve cognitive alignment, and present a research agenda to address these gaps. We argue that cognitive alignment represents a likely impediment to AI adoption in many envisioned applications, and that addressing it is important for creating AI systems on which users are both willing and justified to rely.

A growing body of work in Ethical AI attempts to capture human moral judgments through simple computational models. The key question we address in this work is whether such simple AI models capture the critical nuances of moral decision-making by focusing on the use case of kidney allocation. We conducted twenty interviews where participants explained their rationale for their judgments about who should receive a kidney. We observe participants: (a) value patients' morally-relevant attributes to different degrees; (b) use diverse decision-making processes, citing heuristics to reduce decision complexity; (c) can change their opinions; (d) sometimes lack confidence in their decisions (e.g., due to incomplete information); and (e) express enthusiasm and concern regarding AI assisting humans in kidney allocation decisions. Based on these findings, we discuss challenges of computationally modeling moral judgments as a stand-in for human input, highlight drawbacks of current approaches, and suggest future directions to address these issues.

Recent AI work trends towards incorporating human-centric objectives, with the explicit goal of aligning AI models to personal preferences and societal values. Using standard preference elicitation methods, researchers and practitioners build models of human decisions and judgments, which are then used to align AI behavior with that of humans. However, models commonly used in such elicitation processes often do not capture the true cognitive processes of human decision making, such as when people use heuristics to simplify information associated with a decision problem. As a result, models learned from people's decisions often do not align with their cognitive processes, and can not be used to validate the learning framework for generalization to other decision-making tasks. To address this limitation, we take an axiomatic approach to learning cognitively faithful decision processes from pairwise comparisons. Building on the vast literature characterizing the cognitive processes that contribute to human decision-making, and recent work characterizing such processes in pairwise comparison tasks, we define a class of models in which individual features are first processed and compared across alternatives, and then the processed features are then aggregated via a fixed rule, such as the Bradley-Terry rule. This structured processing of information ensures such models are realistic and feasible candidates to represent underlying human decision-making processes. We demonstrate the efficacy of this modeling approach in learning interpretable models of human decision making in a kidney allocation task, and show that our proposed models match or surpass the accuracy of prior models of human pairwise decision-making.

Preference elicitation frameworks feature heavily in the research on participatory ethical AI tools and provide a viable mechanism to enquire and incorporate the moral values of various stakeholders. As part of the elicitation process, surveys about moral preferences, opinions, and judgments are typically administered only once to each participant. This methodological practice is reasonable if participants' responses are stable over time such that, all other relevant factors being held constant, their responses today will be the same as their responses to the same questions at a later time. However, we do not know how often that is the case. It is possible that participants' true moral preferences change, are subject to temporary moods or whims, or are influenced by environmental factors we don't track. If participants' moral responses are unstable in such ways, it would raise important methodological and theoretical issues for how participants' true moral preferences, opinions, and judgments can be ascertained. We address this possibility here by asking the same survey participants the same moral questions about which patient should receive a kidney when only one is available ten times in ten different sessions over two weeks, varying only presentation order across sessions. We measured how often participants gave different responses to simple (Study One) and more complicated (Study Two) repeated scenarios. On average, the fraction of times participants changed their responses to controversial scenarios was around 10-18% across studies, and this instability is observed to have positive associations with response time and decision-making difficulty. We discuss the implications of these results for the efficacy of moral preference elicitation, highlighting the role of response instability in causing value misalignment between stakeholders and AI tools trained on their moral judgments.

Computational preference elicitation methods are tools used to learn people's preferences quantitatively in a given context. Recent works on preference elicitation advocate for active learning as an efficient method to iteratively construct queries (framed as comparisons between context-specific cases) that are likely to be most informative about an agent's underlying preferences. In this work, we argue that the use of active learning for moral preference elicitation relies on certain assumptions about the underlying moral preferences, which can be violated in practice. Specifically, we highlight the following common assumptions (a) preferences are stable over time and not sensitive to the sequence of presented queries, (b) the appropriate hypothesis class is chosen to model moral preferences, and (c) noise in the agent's responses is limited. While these assumptions can be appropriate for preference elicitation in certain domains, prior research on moral psychology suggests they may not be valid for moral judgments. Through a synthetic simulation of preferences that violate the above assumptions, we observe that active learning can have similar or worse performance than a basic random query selection method in certain settings. Yet, simulation results also demonstrate that active learning can still be viable if the degree of instability or noise is relatively small and when the agent's preferences can be approximately represented with the hypothesis class used for learning. Our study highlights the nuances associated with effective moral preference elicitation in practice and advocates for the cautious use of active learning as a methodology to learn moral preferences.

Real-world applications of machine learning (ML) predictive algorithms often propagate negative stereotypes and social biases against marginalized groups. In response, the field of fair machine learning has proposed technical solutions for various settings that aim to correct the biases in algorithmic predictions. These solutions remove the dependence of the final prediction on the protected attributes (like gender or race) and/or ensure that prediction performance is similar across demographic groups. Yet, recent studies assessing the impact of these solutions in practice demonstrate their ineffectiveness in tackling real-world inequalities. Given this lack of real-world success, it is essential to take a step back and question the design motivations of algorithmic fairness interventions. We use popular legal anti-discriminatory principles, specifically anti-classification and anti-subordination principles, to study the motivations of fairness interventions and their applications. The anti-classification principle suggests addressing discrimination by ensuring that decision processes and outcomes are independent of the protected attributes of individuals. The anti-subordination principle, on the other hand, argues that decision-making policies can provide equal protection to all only by actively tackling societal hierarchies that enable structural discrimination, even if that requires using protected attributes to address historical inequalities. Through a survey of the fairness mechanisms and applications, we assess different components of fair ML approaches from the perspective of these principles. We argue that the observed shortcomings of fair ML algorithms are similar to the failures of anti-classification policies and constitute violations of the anti-subordination principle. Correspondingly, we propose guidelines for algorithmic fairness interventions to adhere to the anti-subordination principle. In doing so, we hope to bridge critical concepts between legal frameworks for nondiscrimination and fairness in machine learning.

In many predictive contexts (e.g., credit lending), true outcomes are only observed for samples that were positively classified in the past. These past observations, in turn, form training datasets for classifiers that make future predictions. However, such training datasets lack information about the outcomes of samples that were (incorrectly) negatively classified in the past and can lead to erroneous classifiers. We present an approach that trains a classifier using available data and comes with a family of exploration strategies to collect outcome data about subpopulations that otherwise would have been ignored. For any exploration strategy, the approach comes with guarantees that (1) all sub-populations are explored, (2) the fraction of false positives is bounded, and (3) the trained classifier converges to a “desired” classifier. The right exploration strategy is context-dependent; it can be chosen to improve learning guarantees and encode context-specific group fairness properties. Evaluation on real-world datasets shows that this approach consistently boosts the quality of collected outcome data and improves the fraction of true positives for all groups, with only a small reduction in predictive utility.

In real-world classification settings, such as loan application evaluation or content moderation on online platforms, individuals respond to classifier predictions by strategically updating their features to increase their likelihood of receiving a particular (positive) decision (at a certain cost). Yet, when different demographic groups have different feature distributions or pay different update costs, prior work has shown that individuals from minority groups often pay a higher cost to update their features. Fair classification aims to address such classifier performance disparities by constraining the classifiers to satisfy statistical fairness properties. However, we show that standard fairness constraints do not guarantee that the constrained classifier reduces the disparity in strategic manipulation cost. To address such biases in strategic settings and provide equal opportunities for strategic manipulation, we propose a constrained optimization framework that constructs classifiers that lower the strategic manipulation cost for minority groups. We develop our framework by studying theoretical connections between group-specific strategic cost disparity and standard selection rate fairness metrics (e.g., statistical rate and true positive rate). Empirically, we show the efficacy of this approach over multiple real-world datasets.

Automatically assigning tasks to people is challenging because human performance can vary across tasks for many reasons. This challenge is further compounded in real-life settings in which no oracle exists to assess the quality of human decisions and task assignments made. Instead, we find ourselves in a "closed" decision-making loop in which the same fallible human decisions we rely on in practice must also be used to guide task allocation. How can imperfect and potentially biased human decisions train an accurate allocation model? Our key insight is to exploit weak prior information on human-task similarity to bootstrap model training. We show that the use of such a weak prior can improve task allocation accuracy, even when human decision-makers are fallible and biased. We present both theoretical analysis and empirical evaluation over synthetic data and a social media toxicity detection task. Results demonstrate the efficacy of our approach.

In hybrid human-machine deferral frameworks, a classifier can defer uncertain cases to human decision-makers (who are often themselves fallible). Prior work on simultaneous training of such classifier and deferral models has typically assumed access to an oracle during training to obtain true class labels for training samples, but in practice there often is no such oracle. In contrast, we consider a"closed"decision-making pipeline in which the same fallible human decision-makers used in deferral also provide training labels. How can imperfect and biased human expert labels be used to train a fair and accurate deferral framework? Our key insight is that by exploiting weak prior information, we can match experts to input examples to ensure fairness and accuracy of the resulting deferral framework, even when imperfect and biased experts are used in place of ground truth labels. The efficacy of our approach is shown both by theoretical analysis and by evaluation on two tasks.

Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL An Anti-Subordination Approach to Fair Classification 27 Pages Posted: 24 May 2022 See all articles by Vijay KeswaniVijay KeswaniYale UniversityL. Elisa CelisYale University Date Written: May 01, 2022 Abstract Many recent studies have pointed out that classification algorithms often propagate negative stereotypes and social biases against historically-oppressed groups. In response, the field of fair machine learning has proposed technical solutions for a variety of settings that aim to correct the biases in the predictions of classification algorithms. These solutions remove the dependence of the final classifier prediction on the protected attributes (like gender or race) and/or ensure that performance is similar across groups defined by the protected attributes. Yet, they may not address the historical and societal inequalities that enable the discrimination in the first place. This can lead to scenarios where the deployment of a fair classification algorithm does not reduce the underlying disparities, and in certain cases, even exacerbates them. Given the possibility of such negative impacts of fair classification, it is important to take a step back and question the motivations of fair classification.We use the legal framework of anti-subordination to study the motivations of a fair classifier and its applications. The anti-subordination principle argues that, in the light of existing and historical biases against disadvantaged groups, decision-making policies can provide equal protection to all groups by actively addressing the historical inequalities faced by disadvantaged groups. Through a survey of the implicit fairness mechanisms in literature, we assess different components of a fair classification system from an anti-subordination perspective. We argue that inappropriate design choices for these components can constitute a violation of the anti-subordination principle, as observed by recent results on negative impacts of algorithmic fairness interventions. Correspondingly, we propose guidelines that a fair machine learning algorithm could follow to adhere to the anti-subordination principle. In doing so, we hope to bridge critical concepts between legal frameworks for non-discrimination and fairness in machine learning. Suggested Citation: Suggested Citation Keswani, Vijay and Celis, L. Elisa, An Anti-Subordination Approach to Fair Classification (May 01, 2022). Available at SSRN: https://ssrn.com/abstract=4116835 Vijay Keswani (Contact Author) Yale University ( email ) New Haven, CT 06520United States L. Elisa Celis Yale University ( email ) 493 College StNew Haven, CT CT 06520United States203.432.0666 (Phone) Download This Paper Open PDF in Browser Do you want regular updates from SSRN on Twitter? Follow SSRN Paper statistics Downloads 14 Abstract Views 39 PlumX Metrics Related eJournals Discrimination, Law & Justice eJournal Follow Discrimination, Law & Justice eJournal Subscribe to this fee journal for more curated articles on this topic FOLLOWERS 1,281 PAPERS 16,618 This Journal is curated by: Francisco Valdes at University of Miami - School of Law Machine Learning eJournal Follow Machine Learning eJournal Subscribe to this fee journal for more curated articles on this topic FOLLOWERS 388 PAPERS 4,266 Artificial Intelligence - Law, Policy, & Ethics eJournal Follow Artificial Intelligence - Law, Policy, & Ethics eJournal Subscribe to this fee journal for more curated articles on this topic FOLLOWERS 288 PAPERS 3,134 This Journal is curated by: J. B. Ruhl at Vanderbilt University - Law School, Dennis D. Hirsch at Ohio State University (OSU) - Michael E. Moritz College of Law Feedback Feedback to SSRN Feedback (required) Email (required) Submit If you need immediate assistance, call 877-SSRNHelp (877 777 6435) in the United States, or +1 212 448 2500 outside of the United States, 8:30AM to 6:00PM U.S. Eastern, Monday - Friday. Submit a Paper Section 508 Text Only Pages SSRN Quick Links SSRN Solutions Research Paper Series Conference Papers Partners in Publishing Jobs & Announcements Newsletter Sign Up SSRN Rankings Top Papers Top Authors Top Organizations About SSRN SSRN Objectives Network Directors Presidential Letter Announcements Contact us FAQs Copyright Terms and Conditions Privacy Policy We use cookies to help provide and enhance our service and tailor content. To learn more, visit Cookie Settings. This page was processed by aws-apollo5 in 0.281 seconds

Elections are the central institution of democratic processes, and often the elected body – in either public or private governance – is a committee of individuals. To ensure the legitimacy of elected bodies, the electoral processes should guarantee that diverse groups are represented, in particular members of groups that are marginalized due to gender, ethnicity, or other socially salient attributes. To address this challenge of representation, we have designed a novel participatory electoral process coined the Representation Pact, implemented with the support of a computational system. That process explicitly enables voters to flexibly decide on representation criteria in a first round, and then lets them vote for candidates in a second round. After the two rounds, a counting method is applied, which selects the committee of candidates that maximizes the number of votes received in the second round, conditioned on satisfying the criteria provided in the first round. With the help of a detailed use case that applied this process in a primary election of 96 representatives in Switzerland, we explain how this method contributes to fairness in political elections by achieving a better "descriptive representation". Further, based on this use case, we identify lessons learnt that are applicable to participatory computational systems used in societal or political contexts. Good practices are identified and presented.

We study a variant of a recently introduced min-max optimization framework where the max-player is constrained to update its parameters in a greedy manner until it reaches a first-order stationary point. Our equilibrium definition for this framework depends on a proposal distribution which the min-player uses to choose directions in which to update its parameters. We show that, given a smooth and bounded nonconvex-nonconcave objective function, access to any proposal distribution for the min-player's updates, and stochastic gradient oracle for the max-player, our algorithm converges to the aforementioned approximate local equilibrium in a number of iterations that does not depend on the dimension. The equilibrium point found by our algorithm depends on the proposal distribution, and when applying our algorithm to train GANs we choose the proposal distribution to be a distribution of stochastic gradients. We empirically evaluate our algorithm on challenging nonconvex-nonconcave test-functions and loss functions arising in GAN training. Our algorithm converges on these test functions and, when used to train GANs, trains stably on synthetic and real-world datasets and avoids mode collapse

Machine learning models are often implemented in cohort with humans in the pipeline, with the model having an option to defer to a domain expert in cases where it has low confidence in its inference. Our goal is to design mechanisms for ensuring accuracy and fairness in such prediction systems that combine machine learning model inferences and domain expert predictions. Prior work on "deferral systems" in classification settings has focused on the setting of a pipeline with a single expert and aimed to accommodate the inaccuracies and biases of this expert to simultaneously learn an inference model and a deferral system. Our work extends this framework to settings where multiple experts are available, with each expert having their own domain of expertise and biases. We propose a framework that simultaneously learns a classifier and a deferral system, with the deferral system choosing to defer to one or more human experts in cases of input where the classifier has low confidence. We test our framework on a synthetic dataset and a content moderation dataset with biased synthetic experts, and show that it significantly improves the accuracy and fairness of the final predictions, compared to the baselines. We also collect crowdsourced labels for the content moderation task to construct a real-world dataset for the evaluation of hybrid machine-human frameworks and show that our proposed framework outperforms baselines on this real-world dataset as well.

We present an optimization framework for learning a fair classifier in the presence of noisy perturbations in the protected attributes. Compared to prior work, our framework can be employed with a very general class of linear and linear-fractional fairness constraints, can handle multiple, non-binary protected attributes, and outputs a classifier that comes with provable guarantees on both accuracy and fairness. Empirically, we show that our framework can be used to attain either statistical rate or false positive rate fairness guarantees with a minimal loss in accuracy, even when the noise is large, in two real-world datasets.

Discussions on Twitter involve participation from different communities with different dialects and it is often necessary to summarize a large number of posts into a representative sample to provide a synopsis. Yet, any such representative sample should sufficiently portray the underlying dialect diversity to present the voices of different participating communities representing the dialects. Extractive summarization algorithms perform the task of constructing subsets that succinctly capture the topic of any given set of posts. However, we observe that there is dialect bias in the summaries generated by common summarization approaches, i.e., they often return summaries that under-represent certain dialects. The vast majority of existing “fair” summarization approaches require socially salient attribute labels (in this case, dialect) to ensure that the generated summary is fair with respect to the socially salient attribute. Nevertheless, in many applications, these labels do not exist. Furthermore, due to the ever-evolving nature of dialects in social media, it is unreasonable to label or accurately infer the dialect of every social media post. To correct for the dialect bias, we employ a framework that takes an existing text summarization algorithm as a blackbox and, using a small set of dialect-diverse sentences, returns a summary that is relatively more dialect-diverse. Crucially, this approach does not need the posts being summarized to have dialect labels, ensuring that the diversification process is independent of dialect classification/identification models. We show the efficacy of our approach on Twitter datasets containing posts written in dialects used by different social groups defined by race or gender; in all cases, our approach leads to improved dialect diversity compared to standard text summarization approaches.

The availability of large datasets and massive computing power has led to a surge in use of automated decision-making frameworks in a variety of domains; the list of applications includes (and is not limited to) healthcare, risk assessment, advertising, lending, and content moderation. However, the presence of biased datasets and/or use of misrepresentative models has had an impact on these automated frameworks as well; multiple studies have pointed out that prediction algorithms often propagate biases and negative stereotypes against historically-oppressed groups. Given the importance of the applications where automation is being employed currently and the possible lack of human oversight in their deployment, it is important that the learned algorithms not only stop the propagation of such biases but also make efforts in curbing the factors that enable such discrimination in first place.

Assessing the diversity of a dataset of information associated with people is crucial before using such data for downstream applications. For a given dataset, this often involves computing the imbalance or disparity in the empirical marginal distribution of a protected attribute (e.g. gender, dialect, etc.). However, real-world datasets, such as images from Google Search or collections of Twitter posts, often do not have protected attributes labeled. Consequently, to derive disparity measures for such datasets, the elements need to hand-labeled or crowd-annotated, which are expensive processes. We propose a cost-effective approach to approximate the disparity of a given unlabeled dataset, with respect to a protected attribute, using a control set of labeled representative examples. Our proposed algorithm uses the pairwise similarity between elements in the dataset and elements in the control set to effectively bootstrap an approximation to the disparity of the dataset. Importantly, we show that using a control set whose size is much smaller than the size of the dataset is sufficient to achieve a small approximation error. Further, based on our theoretical framework, we also provide an algorithm to construct adaptive control sets that achieve smaller approximation errors than randomly chosen control sets. Simulations on two image datasets and one Twitter dataset demonstrate the efficacy of our approach (using random and adaptive control sets) in auditing the diversity of a wide variety of datasets.