Investigates fairness in pipeline models where individuals may drop out.
problem Fairness in pipeline models where individuals may drop out and subsequent stages depend on remaining individuals.
method Rigorous framework for evaluating fairness guarantees, showing that naïve auditing is insufficient and dependence must exist between stages.
result Fairness in pipelines can be arbitrary, even with just two stages, and requires dependence between stages.
This work facilitates ensuring fairness of machine learning in the real world by decoupling fairness considerations in compound decisions. In particular, this work studies how fairness propagates through a compound decision-making processes, which we call a pipeline. Prior work in algorithmic fairness only focuses on f…
Fair Adversarial Networks remove bias from data.
problem Bias in datasets leads to biased outcomes, making analysis illegal and sub-optimal.
method Remove bias from data by altering all proxy variables, ensuring fairness without changing analytical pipelines.
result Fair Adversarial Networks effectively remove bias from data.
The paper aims to improve reproducibility of machine learning experiments.
problem Lack of reproducibility in machine learning experiments.
method Investigates factors affecting reproducibility and proposes FAIR data practices.
result Preliminary results show the role of ProvBook in capturing and comparing provenance.
While harms of allocation have been increasingly studied as part of the subfield of algorithmic fairness, harms of representation have received considerably less attention. In this paper, we formalize two notions of stereotyping and show how they manifest in later allocative harms within the machine learning pipeline. …
Proposes a method to achieve quantile fairness in predictions.
problem Lack of research on quantile fairness in socially sensitive domains.
method Introduces a framework to learn a real-valued quantile function under Demographic Parity fairness.
result Demonstrates superior empirical performance and uncovering fairness-accuracy trade-offs.
New approach to fairness in machine learning models using conformal prediction.
problem Fairness in machine learning models' downstream decision-making.
method Theoretical derivation and empirical evaluation of label-clustered conformal prediction.
result Label-clustered conformal prediction often provides a favorable balance between utility and substantive fairness.
The paper explores fairness in credit scoring using machine learning.
problem The lack of research on fair machine learning in credit scoring.
method Revisits statistical fairness criteria, catalogs algorithmic options, and empirically compares fairness processors.
result Multiple fairness criteria can be approximately satisfied at once, and fair processors deliver a good balance between profit and fairness.
Integrates differential privacy and demographic parity in multi-class classification.
problem Ensuring fairness and privacy in sensitive applications.
method Designs DP2DP algorithm that enforces both demographic parity and differential privacy.
result DP2DP converges towards demographic parity at nearly the same rate as non-private methods, achieving state-of-the-art trade-offs.
Benchmark assesses fairness in algorithmic uncertainty, revealing consistent and calibrated estimates improve fairness.
problem Challenges in managing uncertainty in fairness evaluations for predictive algorithms.
method Introduces FairlyUncertain, an axiomatic benchmark for evaluating uncertainty in fairness.
result Consistent and calibrated uncertainty estimates improve fairness without explicit fairness interventions.
Modyn automates continuous ML model training on growing datasets.
problem Continuous model retraining is costly and impractical.
method Data-centric ML platform with policies for continuous training.
result Modyn enables high throughput training with sample-level data selection.
This paper tackles hidden technical debts in fair ML systems for Fintech.
problem Building fair machine learning systems in financial services.
method Examining key stages of ML system development and deployment.
result Technical debts exist in deploying fair ML systems in Fintech.
The study diagnoses fairness issues in healthcare models under distribution shifts.
problem Understanding and diagnosing fairness changes in machine learning models under distribution shifts in healthcare.
method Causal framing and conditional independence tests to characterize distribution shifts.
result Knowledge of distribution shifts helps diagnose fairness transfer failures, including complex cases.
We propose a fair principal component analysis method that balances reconstruction error and subgroup fairness.
problem Fairness and robustness in principal component analysis for consequential domains.
method Distributionally robust optimization over the Stiefel manifold with a Riemannian subgradient descent.
result The proposed method achieves better performance on real-world datasets compared to state-of-the-art baselines.
ADMM solves constrained CASH problems by breaking them into smaller, manageable pieces.
problem Handling black-box constraints in CASH problems.
method Leverages ADMM optimization framework to decompose CASH problems.
result ADMM facilitates incorporation of black-box constraints.
Unified framework for fair classification with group-blindness/awareness guarantees.
problem Challenges in enforcing fairness and group-blindness in binary classification.
method Unified framework based on post-processing procedure, applicable to various group fairness notions.
result Minimax rate-optimality of the proposed algorithm with controlled excess risk.
The study quantifies and compares aleatoric and epistemic discrimination in ML models.
problem Sources of discrimination in ML models and their impact on performance.
method Quantifying aleatoric and epistemic discrimination using statistical experiments and model accuracy.
result State-of-the-art fairness interventions are effective at removing epistemic discrimination but not aleatoric discrimination in datasets with missing values.
Proposes a non-adversarial method for distribution matching.
problem Stability and optimization challenges in adversarial matching methods.
method Non-adversarial VAE-based matching method with alignment upper bounds.
result Demonstrates applicability of non-adversarial matching methods without modifying original architectures.
Improves accuracy and fairness in prediction systems with multiple domain experts.
problem Designing unbiased and accurate deferral systems with multiple experts.
method Proposes a framework for learning a classifier and deferral system that chooses to defer to multiple human experts.
result Significantly improves accuracy and fairness of final predictions compared to baselines.
New fairness criterion for risk-sensitive decisions in regulated industries.
problem Ensuring equitable outcomes in risk-sensitive decision-making.
method Marginal fairness for generalized distortion risk measures, two-step decision-making process.
result Ensures fairness in decision-making under risk measures, regardless of protected attributes.
Multiverse analysis helps prevent fairness hacking and evaluate model design decisions.
problem Downstream effects of ADM systems depend on implicit design and evaluation decisions.
method Turn implicit decisions into explicit ones, create a grid of decision combinations, compute fairness and performance metrics.
result Decisions regarding evaluation can lead to vastly different fairness metrics for the same model.
Machine learning models are widely adopted in scenarios that directly affect people. The development of software systems based on these models raises societal and legal concerns, as their decisions may lead to the unfair treatment of individuals based on attributes like race or gender. Data preparation is key in any ma…
New technique reduces gender discrimination in credit lending models.
problem Bias and unfairness in credit lending predictions.
method Subgroup Threshold Optimizer (STO) technique.
result Reduces gender discrimination by over 90%.
Study reveals AI skin cancer classifiers underperform for darker skin phototypes, advocating for fairness auditing.
problem AI bias in dermatology, particularly for darker skin phototypes.
method Predictive Representativity (PR) framework, evaluating classifiers on HAM10000 and BOSQUE Test sets.
result Substantial performance disparities by skin phototype, highlighting AI bias.
Now that machine learning algorithms lie at the center of many resource allocation pipelines, computer scientists have been unwittingly cast as partial social planners. Given this state of affairs, important questions follow. What is the relationship between fairness as defined by computer scientists and notions of soc…
This study simulates biases in classifiers to assess fairness.
problem Mitigating biases in predictive models to ensure fairness.
method Agent-based model (ABM) to generate synthetic datasets with controlled biases, applied to offline and online learning approaches.
result Demonstrates how biases in data affect classifier outcomes and how mitigations impact feature usage.
FST.ai 2.0 improves Taekwondo decision-making with AI, reducing review time and increasing trust.
problem Fair, transparent, and explainable decision-making in Taekwondo.
method Pose-based action recognition, epistemic uncertainty modeling, interactive dashboards.
result 85% reduction in decision review time, 93% referee trust in AI-assisted decisions.
MARS-Gym framework for marketplaces to train and evaluate recommender systems.
problem Challenges in designing, training, and evaluating recommender systems in marketplaces.
method Open-source framework for Reinforcement Learning agents in marketplaces.
result Empowers researchers and engineers to quickly build and evaluate agents for recommendations.
Open dataset and pipeline for realistic OPE research.
problem Lack of realistic and reproducible OPE experimental studies.
method Public logged bandit dataset and Python software.
result Enables experimental comparisons of OPE estimators.
The persistence of racial inequality in the U.S. labor market against a general backdrop of formal equality of opportunity is a troubling phenomenon that has significant ramifications on the design of hiring policies. In this paper, we show that current group disparate outcomes may be immovable even when hiring decisio…
This work addresses fairness in ML models by training and evaluating attribute classifiers under uncertain and incomplete data.
problem Challenges in fairness metrics due to uncertain and incomplete data.
method Developed a theoretical and empirical analysis to understand and improve bias estimation in the data-scarce regime.
result The test accuracy of the attribute classifier is not always correlated with its effectiveness in bias estimation.
Complex statistical machine learning models are increasingly being used or considered for use in high-stakes decision-making pipelines in domains such as financial services, health care, criminal justice and human services. These models are often investigated as possible improvements over more classical tools such as r…
AVATAR uses a surrogate model to quickly evaluate ML pipelines, saving time and resources.
problem Time-consuming evaluation of ML pipelines limits exploration of complex models.
method AVATAR employs a surrogate model to assess pipeline validity without execution.
result AVATAR accelerates ML pipeline evaluation, improving efficiency in complex scenarios.
The paper addresses bias amplification in prediction and decision-making using causal analysis.
problem Bias amplification in automated systems, especially after thresholding.
method Introduces margin complement and causal decomposition of prediction disparities.
result Disparity in predictor Y ^ \widehat Y Y can be decomposed into causal influences of X X X on S S S and M M M . Hyperparameter tuning of multi-stage pipelines introduces a significant computational burden. Motivated by the observation that work can be reused across pipelines if the intermediate computations are the same, we propose a pipeline-aware approach to hyperparameter tuning. Our approach optimizes both the design and exe…
The constant introduction of standardized benchmarks in the literature has helped accelerating the recent advances in meta-learning research. They offer a way to get a fair comparison between different algorithms, and the wide range of datasets available allows full control over the complexity of this evaluation. Howev…
Classical Machine Learning (ML) pipelines often comprise of multiple ML models where models, within a pipeline, are trained in isolation. Conversely, when training neural network models, layers composing the neural models are simultaneously trained using backpropagation. We argue that the isolated training scheme of ML…
Study proposes a statistical testing framework for evaluating clustering pipelines.
problem Quantifying the statistical reliability of clustering results from data analysis pipelines.
method Selective inference-based statistical testing framework for clustering pipelines.
result The proposed test controls the type I error rate and is effective in validating clustering results.
MLModelScope streamlines ML/DL model evaluation and benchmarking.
problem Challenges in evaluating and benchmarking ML/DL models.
method Open-source, framework/hardware agnostic platform with distributed design.
result Demonstrates the impact of model evaluation pipelines and HW/SW choices.
Study proves convergence of subgradients for optimal transport-based objectives.
problem Ensuring statistical consistency and optimization stability in transport-based models.
method Proves graphical convergence of subdifferentials to the subdifferential of the population objective.
result Standard subgradient methods consistently approach stationary points of the population-level problem.
Two simple methods learn fair metrics from data to improve fairness in ML tasks.
problem Lack of widely accepted fair metrics for many ML tasks hinders individual fairness adoption.
method Presented two simple ways to learn fair metrics from various data types.
result Fair training with learned metrics improves fairness on three ML tasks.
New concept of within-group fairness improves AI fairness without sacrificing accuracy.
problem Fairness issues in AI models treating individuals in the same sensitive group unfairly.
method Introducing within-group fairness, proposing mathematical definitions, and developing learning algorithms.
result Improves within-group fairness without sacrificing accuracy and between-group fairness.
A new fairness metric for decision-making algorithms, conditioning on known fair variables.
problem Fairness issues in decision-making systems.
method Conditional fairness metric, Derivable Conditional Fairness Regularizer (DCFR), adversarial representation.
result Traditional fairness notations are special cases of the new conditional fairness notation.
Data science relies on pipelines that are organized in the form of interdependent computational steps. Each step consists of various candidate algorithms that maybe used for performing a particular function. Each algorithm consists of several hyperparameters. Algorithms and hyperparameters must be optimized as a whole …
Fair Mixup improves fairness in classifiers by interpolating between groups.
problem Ensuring fairness in classifiers during training and evaluation.
method Fair Mixup uses interpolation of samples between groups to enforce fairness constraints.
result Fair Mixup ensures better generalization of fairness in various benchmarks.
Paper proposes a modified fairness constraint to address shortcomings of counterfactual fairness.
problem Counterfactual fairness is not a necessary condition for algorithmic fairness.
method Analyzed hypothetical scenario and explicated discrimination to develop causal relevance fairness.
result Causal relevance fairness is a modified constraint that circumvents shortcomings of counterfactual fairness.
DFL framework improves action and outcome fairness in policy learning.
problem Fairness in policy learning, especially action and outcome fairness.
method Integrates action and outcome fairness into a multi-objective optimization problem using a lexicographic weighted Tchebyshev method.
result DFL framework improves both action and outcome fairness with minimal value reduction.
Automates supervised learning pipeline design with matrix and tensor factorization.
problem Designing effective supervised learning pipelines with many choices.
method Uses matrix and tensor factorization to model pipeline search space and develops greedy experiment design protocols.
result Demonstrates the effectiveness of the approach on real-world classification problems.