AEQUITAS automatically tests and improves fairness of machine learning models.
problem Ensuring fairness in machine learning models used in sensitive domains.
method Probabilistic search over input space to discover discriminatory inputs, leveraging robustness of models.
result AEQUITAS effectively generates inputs to uncover and improve fairness in machine learning models.
Transfer learning can worsen fairness, study finds.
problem Transfer learning can reduce fairness in predictions.
method Examined fairness of standard transfer and multi-task learning algorithms.
result Both standard algorithms suffer from discriminatory transfer.
Predictive models are increasingly deployed for the purpose of determining access to services such as credit, insurance, and employment. Despite potential gains in productivity and efficiency, several potential problems have yet to be addressed, particularly the potential for unintentional discrimination. We present an…
The study compares uniform-price and discriminatory auctions in terms of learning difficulty.
problem Comparing the learning difficulty of uniform-price and discriminatory multi-unit auctions.
method Characterization of learning difficulty through regret minimization in both full-information and bandit feedback settings.
result Regret scales similarly for both auction formats under full-information, but uniform-price auctions can achieve faster learning rates.
This paper defines less discriminatory algorithms and explores their feasibility.
problem Creating algorithms that are less discriminatory while meeting business needs.
method Formal definition of less discriminatory algorithms, evaluation of feasibility, and search for alternatives.
result Formal definitions of less discriminatory algorithms face challenges due to lack of held-out data, necessitating a reliance on reasonableness standards.
Private release of sensitive data enables fair learning.
problem Learning fair predictors with restricted sensitive data.
method Private release of sensitive demographic data, adapting non-discriminatory learners.
result The approach provides theoretical guarantees on performance for fair predictors.
Many datasets can be viewed as a noisy sampling of an underlying space, and tools from topological data analysis can characterize this structure for the purpose of knowledge discovery. One such tool is persistent homology, which provides a multiscale description of the homological features within a dataset. A useful re…
Paper detects proxies in linear regression models causing discrimination.
problem Discrimination in machine learning models using proxies for protected attributes.
method Formulated a definition of proxy use, identified proxies via second-order cone program, and extended to justified business necessity.
result Proxies in linear regression models can be efficiently identified and removed to reduce discrimination.
Procedure for determining less discriminatory alternatives in AI audits with limited resources.
problem Difficulty in proving less discriminatory alternatives in AI audits due to resource constraints.
method Closed-form upper bound for loss-fairness Pareto frontier, enabling claimants to fit PFs without training large models.
result A scaling law for loss-fairness Pareto frontiers, allowing claimants to determine if an LDA exists with limited resources.
Whole MILC learns brain disorder dynamics from unlabeled data.
problem Learning spatio-temporal brain disorder dynamics from unlabeled data.
method Self-supervised pre-training of whole MILC on unlabeled healthy control data.
result Whole MILC outperforms existing methods and provides diagnostic insights.
Paper introduces a novel method to generate diverse inputs for neural programming by example.
problem Synthesizing programs from input/output pairs using machine learning.
method Uses an SMT solver to generate diverse input-output pairs.
result Generated inputs improve model performance and generalization.
Deep ensemble learning framework approximates any functions from input to output space.
problem Approximating any functions from input to output space using deep learning models.
method Proposes a deep ensemble learning framework that combines the results of multiple unit models to achieve universal approximation of functions.
result The deep ensemble learning framework can achieve a universal approximation of any functions from the input space to the output space when the unit model mappings are bounded, sigmoidal, and discriminatory.
r-STSF improves TSC accuracy and interpretability.
problem Lack of interpretability in state-of-the-art TSC methods.
method Randomized-Supervised Time Series Forest (r-STSF) using interval-based approach and ensemble of randomized trees.
result r-STSF achieves state-of-the-art accuracy and enables interpretability.
Paper exposes vulnerabilities in interpreting machine learning models using adversarial attacks on PD plots.
problem Vulnerability of permutation-based interpretation methods, particularly PD plots, to adversarial attacks.
method Adversarial framework to manipulate black-box models and produce deceptive PD plots.
result It is possible to hide discriminatory behaviors in machine learning models through interpretation tools like PD plots.
This article guides data scientists on avoiding discrimination in machine learning.
problem Machine learning systems can create or exacerbate societal disparities.
method Provides a taxonomy of practices and measures to mitigate discrimination.
result Data scientists should be intentional about modeling and reducing discriminatory outcomes.
The intention with this paper is to provide all the estimation concepts and techniques that are needed to implement a two-phases approach to the parametric estimation of probability of default (PD) curves. In the first phase of this approach, a raw PD curve is estimated based on parameters that reflect discriminatory p…
New methods explain NE embeddings by identifying key variables.
problem Lack of interpretability in NE techniques.
method Combining PCA, Q-residuals, Hotelling's T2, and visualization.
result Identifies discriminatory features not seen in standard approaches.
The paper examines the stability of binary choice models using Gini index and scoring indicators.
problem Stability and discriminatory power of binary choice models.
method Derives the real Gini index and incorporates PSI and KS statistics into the model.
result The real Gini index should be less than the calculated Gini index when the population distribution changes.
New metric MADD assesses fairness of predictive student models.
problem Predictive student models can be biased and unfair, leading to discrimination.
method Proposes MADD metric to analyze model's discriminatory behaviors.
result Fair predictive performance does not guarantee fair behaviors or outcomes.
Expands Bayesian experiment design framework to account for model discrepancies.
problem Model misspecification in Bayesian optimal experiment design.
method Introduces Expected General Information Gain and Expected Discriminatory Information criteria.
result Demonstrates improved robustness and detection capabilities in experiment design.
Efficient methods estimate concordance probability for big data.
problem Efficiently calculating concordance probability in large datasets.
method Proposes two estimation methods for discrete and continuous settings.
result Estimators are accurate and computationally efficient.
LDA-XGB1 balances fairness and accuracy in lending models.
problem Fair lending practices and model interpretability in binary classification.
method Biobjective optimization using binning and information value, leveraging XGBoost.
result Achieves effective balance between accuracy, fairness, and interpretability.
Discriminatory trade liberalization policies are becoming more popular among world economies. Countries are motivated to enter for regional trade agreements to capture faster economic growth for alleviating poverty. In developing economies like most of the member countries of the Association of South East Asian Nations…
Critical initialisation strategies are identified for noisy ReLU networks.
problem Understanding signal propagation in noisy rectifier neural networks.
method Developed a new framework for signal propagation in stochastic regularized neural networks, incorporating various noise distributions.
result Critical initialisation strategies for multiplicative noise (e.g. dropout) are identified, but not for additive noise.
The paper tackles fairness in supervised learning using information theory.
problem Discrimination in decision rules derived from biased historical data.
method Information theoretic framework for designing fair predictors, using equalized odds criterion.
result Designing predictors that are independent of a sensitive attribute while generalizing well.
Remote explainability is impossible for single explanations, showing discriminatory features.
problem Remote explainability for machine learning models is challenging due to the lack of transparency.
method Analogy with club bouncer and proof of impossibility of remote explainability for single explanations.
result Remote explainability for single explanations is impossible, as shown by an attack that hides discriminatory features.
Study evaluates fairness metrics in biased datasets.
problem Detecting bias in machine learning models trained on biased data.
method Causal inference with observational data, investigating six fairness metrics.
result Best practice guidelines for selecting fairness metrics.
New measure mitigates algorithmic discrimination in rich class of computations.
problem Discrimination in algorithmic predictions due to data analysis biases.
method Develops multicalbration, a new measure of algorithmic fairness.
result Multicalibration guarantees accurate predictions for every subpopulation.
Proposes a method for evaluating multiple dimensions of organizational effectiveness using DEA.
problem Evaluating multiple dimensions of organizational effectiveness in large data sets.
method Introduces two regularized DEA models (SBM and GP-SBM) to estimate both dimension-specific and aggregate efficiency scores.
result Demonstrates improved efficiency and validity compared to conventional methods.
FAE framework tackles fairness in machine learning by balancing data and adjusting decision boundaries.
problem Discrimination in automated decision-making based on machine learning algorithms.
method Combines pre- and post-processing fairness interventions to address group imbalance, class imbalance, and class overlap.
result Improves fairness in machine learning models by balancing data and adjusting decision boundaries.
A new framework separates classifier calibration and discrimination.
problem Combining reliability and resolution in probabilistic predictions.
method Manokhin Probability Matrix separates reliability and resolution using Spiegelhalter Z-statistic and AUC-ROC.
result Classifiers are categorized into four archetypes: Eagle, Bull, Sloth, and Mole.
A new algorithm for efficiently removing specific classes from a model without retraining.
problem Class forgetting in machine learning models.
method Estimating retain and forget spaces using SVD, removing shared information, and updating weights.
result Achieved up to 1.38% accuracy improvement on ImageNet dataset with minimal samples.
Recidivism prediction instruments provide decision makers with an assessment of the likelihood that a criminal defendant will reoffend at a future point in time. While such instruments are gaining increasing popularity across the country, their use is attracting tremendous controversy. Much of the controversy concerns …
A new method identifies class-specific covariates in multi-class prediction tasks.
problem Identifying covariates specifically associated with one or more outcome classes in multi-class prediction tasks.
method Introducing multi forests (MuFs) with multi-way and binary splits to measure class-associated discriminatory ability.
result The multi-class VIM specifically ranks class-associated covariates highly, unlike conventional VIMs.
This paper elaborates on the validation requirements for rating systems and probabilities of default (PDs) which were introduced with the New Capital Standards (Basel II). We start in Section 2 with some introductory remarks on the topics and approaches that will be discussed later on. Then we have a view on the develo…
Unified platform SOCRATES for neural network analysis.
problem Analyzing neural networks for bugs and fairness.
method Standardized format, assertion language, and multiple analysis algorithms.
result Unified platform for neural network analysis.
A new divergence measure for distributions with different supports.
problem Divergence not defined for distributions with different supports.
method Define Spread Divergence on modified distributions and maximize discriminatory power.
result Spread Divergence can be used to train generative models.
New clustering method considers causal fairness to avoid bias.
problem Clustering algorithms can unintentionally propagate unfair disparities.
method Integrates causal fairness metrics into clustering algorithms.
result Demonstrates efficacy on datasets with known unfair biases.
Adapts concordance probability for large non-life insurance datasets.
problem Capturing discriminatory ability in large non-life insurance datasets.
method Adapts C-index definition and presents two estimation procedures.
result Validates the new procedures for various versions of C-index.
Recidivism prediction instruments (RPI's) provide decision makers with an assessment of the likelihood that a criminal defendant will reoffend at a future point in time. While such instruments are gaining increasing popularity across the country, their use is attracting tremendous controversy. Much of the controversy c…
We find the optimal error for a constrained regression model under a linear model.
problem Minimizing error while adhering to demographic parity constraints.
method Proposed a minimax optimal error analysis for a demographic parity-constrained regression problem within a linear model.
result The minimax optimal error is characterized by $Θ( rac{dM}{n})$.
This paper examines confidence intervals for class prevalences in shifted datasets.
problem Estimating class prevalences in shifted datasets and distinguishing between confidence and prediction intervals.
method Simulation study comparing different methods for constructing confidence and prediction intervals.
result Discriminatory power of the classifier affects the accuracy of class prevalence estimates.
Deep neural networks have been developed drawing inspiration from the brain visual pathway, implementing an end-to-end approach: from image data to video object classes. However building an fMRI decoder with the typical structure of Convolutional Neural Network (CNN), i.e. learning multiple level of representations, se…
New fair-by-design model reduces bias in recidivism prediction.
problem Discriminatory bias in recidivism prediction models.
method Prototype-based, locally learned, data distribution extraction.
result Reduces bias and provides interpretable rules.
Machine learning can impact people with legal or ethical consequences when it is used to automate decisions in areas such as insurance, lending, hiring, and predictive policing. In many of these scenarios, previous decisions have been made that are unfairly biased against certain subpopulations, for example those of a …
Paper proposes fair ML predictors to avoid discrimination.
problem Discrimination in ML predictors from historical data.
method Proposes two algorithms to adjust ML predictors for fairness.
result Proves fair EO and AA predictors are optimal in performance.
fairadapt uses causal inference to mitigate algorithmic bias in data pre-processing.
problem Mitigating algorithmic bias in machine learning predictions.
method Causal graphical model and observed data to address counterfactual questions.
result The method can help eliminate discrimination and justify fair decisions.
New algorithm reduces discrimination in predictions.
problem Tackles potential discrimination in AI predictions.
method Integrates fairness adjustments into tree-building process.
result Reduces discriminatory predictions without significant loss in accuracy.