New framework shows algorithmic recourse can be harmful.
problem Risks of providing algorithmic recourse in binary classification.
method Abstract learning-theoretic framework comparing risks with and without recourse.
result Providing recourse can be harmful, increasing class uncertainty and mistakes.
COMRECGC finds common recourse for global counterfactual explanations in GNNs.
problem Finding common recourse for global counterfactual explanations in GNNs.
method Formalized the common recourse explanation problem and designed COMRECGC algorithm.
result COMRECGC outperforms strong baselines on four real-world graph datasets.
Recourse explanations can become invalid if collective actions change statistical data.
problem Recourse explanations may become invalid due to collective behavior changing data statistics.
method Formal characterization of conditions under which recourse explanations remain valid under performativity.
result Recourse actions may become invalid if they are influenced by or intervene on non-causal variables.
Paper proposes using pairwise feature comparisons to infer modification costs for user recourse.
problem Learning and inferring user preferences for modifying features in black-box models.
method Bradley-Terry model for inferring feature-wise costs from non-exhaustive human comparison surveys.
result Non-exhaustive human surveys can efficiently learn feature costs, enabling recourse finding.
Develops a new approach for algorithmic recourse in AI systems.
problem Tackles the problem of providing recommendations for reversing negative AI decisions.
method Introduces a causal framework that models recourse as a process over pre- and post-intervention outcomes, allowing for partial stability and resampling of latent variables.
result Demonstrates the value of the proposed methods on real and semi-synthetic datasets.
AReS framework provides interpretable recourses for entire populations.
problem Ensuring meaningful and non-discriminatory recourses for high-stakes decision-making.
method Model agnostic framework for global counterfactual explanations, optimizing correctness and interpretability while minimizing costs.
result AReS enables learning compact rule sets for recourses across subpopulations, with optimality guarantees.
New approach to meaningful and robust algorithmic recourse.
problem Ineffective and unmeaningful algorithmic recourse explanations.
method Meaningful Algorithmic Recourse (MAR) and Effective Algorithmic Recourse (EAR).
result Proposes new constraints for algorithmic recourse that improve both prediction and target.
New probabilistic approaches offer recourse recommendations even when causal models are imperfect.
problem Limited causal knowledge makes guaranteeing algorithmic recourse impossible.
method Two probabilistic approaches: Bayesian model averaging and average effect computation.
result Probabilistic approaches lead to more reliable recourse recommendations.
Proposes a new method for algorithmic recourse in confounded settings.
problem Provides actionable recommendations for individuals affected by automated decisions.
method Relaxes assumptions of no hidden confounding and additive noise, requiring only causal graph and confounding structure.
result Bounds the expected counterfactual effect of recourse actions, ensuring favourable outcomes in expectation.
New findings show local attributions can't be both robust and provide recourse.
problem Ensuring machine learning systems are accountable and provide actionable recourse options.
method Formal definition of recourse sensitivity and counterexamples for popular attribution methods.
result It is impossible for any single attribution method to be both robust and provide recourse.
Improves algorithmic recourse to guide towards both acceptance and improvement.
problem Algorithmic recourse recommendations may not lead to improvement.
method Improvement-Focused Causal Recourse (ICR) requires recommendations to guide towards improvement and leverages causal knowledge to design accurate decision systems.
result ICR guides towards both acceptance and improvement given correct causal knowledge.
Improves global counterfactual explanations for model recourse.
problem Inability to provide explanations beyond local instances.
method Investigates and improves Actionable Recourse Summaries (AReS) for global counterfactual explanations.
result Develops more efficient and interactive explainability tools.
New fairness criteria for algorithmic recourse actions that consider causal relationships.
problem Fairness of recourse actions in algorithmic classification.
method Proposes two new fairness criteria at group and individual levels, explicitly accounting for causal relationships.
result Fairness of recourse is complementary to fairness of prediction, and can be enforced by altering the classifier.
Proposes a new algorithm for accurate tree-based models with guaranteed recourse actions.
problem Ensuring recourse actions for models optimized for predictive performance.
method Formulates and solves an optimization task to ensure reasonable actions for as many instances as possible, using adversarial training techniques.
result Successfully provided reasonable actions to more instances than baselines without significantly degrading accuracy and computational efficiency.
Proposes a new method to explain model predictions for consumer recourse.
problem Current explanation methods fail to provide meaningful recourse to decision subjects.
method Develops feature responsiveness scores to highlight actionable features.
result Standard practices can undermine decision subjects by highlighting unresponsive features.
The paper proposes a method to measure fairness through equality of effort using algorithmic recourse.
problem Measuring fairness through equality of effort in automated systems.
method Applying algorithmic recourse to quantify equality of effort, overcoming previous limitations.
result An algorithm for assessing equality of effort has been developed and validated.
Proposes sparse local and regional counterfactual rules for robust recourses.
problem Challenges in counterfactual explanations, especially stability, synthesis, and implementation.
method Probabilistic framework using Random Forest to derive sparse local and regional counterfactual rules.
result Effective recourses derived from high-density regions, providing sparse and robust counterfactual rules.
The rise in machine learning-assisted decision-making has led to concerns about the fairness of the decisions and techniques to mitigate problems of discrimination. If a negative decision is made about an individual (denying a loan, rejecting an application for housing, and so on) justice dictates that we be able to as…
This work surveys algorithmic recourse, aiming to clarify definitions and solutions.
problem Providing explanations and recommendations to individuals affected by automated decisions.
method Literature review and unified definitions, formulations, and solutions.
result Unified definitions and solutions for algorithmic recourse.
Machine learning models are increasingly used to automate decisions that affect humans - deciding who should receive a loan, a job interview, or a social service. In such applications, a person should have the ability to change the decision of a model. When a person is denied a loan by a credit score, for example, they…
As machine learning is increasingly used to inform consequential decision-making (e.g., pre-trial bail and loan approval), it becomes important to explain how the system arrived at its decision, and also suggest actions to achieve a favorable decision. Counterfactual explanations -- "how the world would have (had) to b…
Procedure verifies if machine learning models assign fixed predictions that preclude access.
problem Models assign fixed predictions that preclude access to credit and employment.
method Model-agnostic recourse verification with reachable sets.
result Models can inadvertently preclude access by assigning fixed predictions.
New AR framework handles missing values for better recourse actions.
problem Existing AR methods fail with missing values, leading to invalid or costly actions.
method Formulates task using multiple imputation and mixed-integer linear optimization.
result Efficacy of new method demonstrated in experiments.
Machine learning based decision making systems are increasingly affecting humans. An individual can suffer an undesirable outcome under such decision making systems (e.g. denied credit) irrespective of whether the decision is fair or accurate. Individual recourse pertains to the problem of providing an actionable set o…
CounteRGAN generates realistic, actionable counterfactuals for machine learning models.
problem Creating realistic and actionable counterfactuals for machine learning models.
method Applying Residual GANs to improve counterfactual realism and actionability.
result CounteRGAN produces counterfactuals with improved realism and actionability, achieving real-time applicability.
This work improves surrogate models for balancing accuracy and cost in multi-fidelity methods.
problem Balancing accuracy and computational cost in multi-fidelity methods.
method Develops context-aware surrogate models for multi-fidelity importance sampling and Bayesian inverse problems.
result Context-aware surrogate models can lead to runtime speedups of up to one order of magnitude.
We give an entirely geometric proof, without recourse to cellular homology, of the fact that ∂2=0 in the chain complex defined by a handle decomposition of a given manifold. Topological invariance of the resulting `handle homology' is a consequence of Cerf theory.
Study on maximizing submodular functions with limited updates, achieving tight bounds and poly-time algorithms.
problem Online submodular maximization with constant recourse.
method Information-theoretic bounds and poly-time randomized algorithms.
result Achieved tight bounds of 2/3 and 3/4 for general and coverage functions, respectively, with a 0.51 approximation.
New framework for contesting algorithmic decisions, not just explaining them.
problem Helping individuals review and correct erroneous algorithmic decisions.
method Operationalized contestability as a natural complement to explainable AI (XAI), identifying three types of evidence for reversal.
result Existing EU legislation already grants individuals legal rights to contest algorithmic decisions.
In a simplified setting, we show how to price invoice non-recourse factoring taking into account not only the credit worthiness of the debtor but also the assignor's one, together with the default correlation between the two. Indeed, the possible default of the assignor might impact the payoff by means of the bankruptc…
New approach to counterfactual reasoning avoids demographic interventions.
problem Limitations of traditional counterfactual reasoning in AI systems.
method Backtracking counterfactual approach instead of interventional.
result Allows addressing social concerns without demographic interventions.
The usual development of the continuous-time random walk (CTRW) proceeds by assuming that the present is one of the jumping times. Under this restrictive assumption integral equations for the propagator and mean escape times have been derived. We generalize these results to the case when the present is an arbitrary tim…
We present the Integrated Size and Price Optimization Problem (ISPO) for a fashion discounter with many branches. Based on a two-stage stochastic programming model with recourse, we develop an exact algorithm and a production-compliant heuristic that produces small optimality gaps. In a field study we show that a distr…
Proposes an efficient method for ordered counterfactual explanations.
problem Insufficient explanation of perturbation vectors for executing actions.
method Mixed-Integer Linear Optimization (MILP) approach for evaluating and extracting optimal pairs of actions and orders.
result Demonstrated effectiveness of the proposed method on real datasets.
Define an arithmetic variety to be the quotient of a bounded symmetric domain by an arithmetic group. An arithmetic variety is algebraic, and the theorem in question states that when one applies an automorphism of the field of complex numbers to the coefficients of an arithmetic variety the resulting variety is again a…
We improve optimization for data with varying variance.
problem Optimizing data with varying variance.
method Generalized learning and optimization frameworks for data-driven optimization.
result Asymptotic and finite sample guarantees for stochastic programs.
Stablecoins offer efficient settlement but externalize costs and risks.
problem Comparing stablecoins to card networks in retail payments.
method Unified analytical framework (CLEAR) across five dimensions.
result Stablecoins are advantageous in closed-loop and high-friction contexts but structurally disadvantaged as open-loop instruments.
Models for recommender systems show similar results in item availability.
problem Model uncertainty in recommender systems.
method Examined different variations of a model for item availability using predictive multiplicity.
result Most models produce similar results in terms of item availability discrepancy.
We recently showed that the S&P500 stock market index is well described by Tsallis non-extensive statistics and nonlinear Fokker-Planck time evolution. We argued that these results should be applicable to a broad range of markets and exchanges where anomalous diffusion and `heavy' tails of the distribution are present.…
There are no known exact formulas for the valuation of a number of exotic options, and this is particularly true for options under discrete monitoring and for American style options. Therefore, one usually recourses to a Monte Carlo Simulation approach, amongst other numerical methods, to estimate the value of these op…
Optimal recovery framework for non-IID data in Hilbert spaces.
problem Generalization in non-IID data scenarios.
method Optimal recovery perspective, semidefinite programming, kernel ridgeless regression.
result Optimal recovery formula coincides with kernel ridgeless regression in some cases.
By mid 2004, the Basel Committee on Banking Supervision (BCBS) is epected to launch its final recommendations on minimum capital requirements in the banking industry. Although there is the intention to arrive at capital charges which concur with economic intuition, the risk weight formulas proposed by the committee wil…
We revisit the optimal investment and consumption model of Davis and Norman (1990) and Shreve and Soner (1994), following a shadow-price approach similar to that of Kallsen and Muhle-Karbe (2010). Making use of the completeness of the model without transaction costs, we reformulate and reduce the Hamilton-Jacobi-Bellma…
When training large machine learning models with many variables or parameters, a single machine is often inadequate since the model may be too large to fit in memory, while training can take a long time even with stochastic updates. A natural recourse is to turn to distributed cluster computing, in order to harness add…
Recommender systems often rely on models which are trained to maximize accuracy in predicting user preferences. When the systems are deployed, these models determine the availability of content and information to different users. The gap between these objectives gives rise to a potential for unintended consequences, co…
AlphaZero assesses new chess variants for balance and dynamics.
problem Designing engaging and balanced game rules, especially for chess variants.
method Used AlphaZero to learn near-optimal strategies for nine chess variants.
result AlphaZero reveals novel strategic and tactical patterns in chess variants.
We study solutions of the Bogomolny equation on R^2\times S^1$ with prescribed singularities. We show that Nahm transform establishes a one-to-one correspondence between such solutions and solutions of the Hitchin equations on a punctured cylinder with the eigenvalues of the Higgs field growing at infinity in a particu…
This paper explores what causal structures can be distinguished by observational and interventional probing schemes.
problem Identifying causal structures with latent variables using observational and interventional data.
method Investigates the power of different probing schemes (observation vs. intervention) to distinguish causal structures.
result Two causal structures are indistinguishable if they share the same mDAG structure.