BREEDS benchmarks assess model robustness to subpopulation shifts.
problem Measuring model robustness to novel subpopulation shifts.
method Controlled synthesis of realistic distribution shifts using class structure.
result Validated model sensitivity and effectiveness of robustness interventions.
New methods adapt conformal prediction to unknown subpopulation shifts.
problem Failure of conformal prediction under unknown subpopulation shifts.
method Proposes new methods that adapt conformal prediction to unknown subpopulation shifts without explicit subpopulation labels.
result Ensures valid coverage guarantees without explicit knowledge of subpopulation structure.
Improves model robustness to shifts in subpopulations.
problem Poor performance of ML models under data distribution shifts.
method Develops group-aware priors (GAP) over neural network parameters.
result Training with GAP yields state-of-the-art performance.
New method detects novel node categories in graphs with distribution shifts.
problem Detecting novel node categories in graphs with distribution shifts.
method Recall-Constrained Optimization with Selective Link Prediction (RECO-SLIP).
result RECO-SLIP outperforms existing methods in detecting novel node categories.
DORO improves DRO's performance and stability in tasks with subpopulation shift.
problem DRO's poor performance and instability in tasks with subpopulation shift.
method DORO, a refined risk function that prevents overfitting to outliers.
result DORO improves DRO's performance and stability on large modern datasets.
The study examines if enforcing fairness helps mitigate bias due to demographic shifts.
problem Bias in ML models due to underrepresented subpopulations in training data.
method Theoretical and simulation analysis of fairness enforcement in target domains.
result Enforcing fairness can lead to the Bayes model in the target domain under certain conditions.
Proposes a label propagation framework for domain adaptation.
problem Subpopulation shift in machine learning domains.
method Label propagation based on a teacher classifier trained on source domain.
result End-to-end finite-sample guarantees on domain adaptation algorithm.
Paper proposes a new regularization method to prevent model degradation under distribution shifts.
problem Model performance degrades under distribution shifts.
method Supervised contrastive learning with heterogeneous similarity.
result The proposed method outperforms existing regularization methods on benchmark datasets.
The study assesses external validity by evaluating worst-case treatment effects across subpopulations.
problem Underrepresentation of marginalized groups and limited study populations.
method Develops a semiparametrically efficient estimator for worst-case treatment effects (WTE) and uses cross-fitting to guard against brittle findings.
result The proposed framework guards against invalid findings due to unanticipated population shifts.
A common goal in statistics and machine learning is to learn models that can perform well against distributional shifts, such as latent heterogeneous subpopulations, unknown covariate shifts, or unmodeled temporal effects. We develop and analyze a distributionally robust stochastic optimization (DRO) framework that lea…
Unified framework for fairness, robustness, and distribution shifts.
problem Diverse failure modes of machine learning systems.
method Formalizes biases as violations of conditional independence and proves equivalence conditions.
result Equivalent effects of biases in different failure modes under specific conditions.
Proposes a method to ensure low losses across all subpopulations in large datasets.
problem Standard practice of minimizing average loss fails to guarantee low losses across all subpopulations in heterogeneous datasets.
method Convex procedure that controls worst-case performance over all subpopulations of a given size with finite-sample convergence guarantees.
result Empirically, the worst-case procedure learns models that do well against unseen subpopulations.
Syntax designs adaptive trials for subpopulations with potential benefits.
problem Identifying subpopulations with positive treatment effects in diverse patient populations.
method Adaptive patient recruitment and synthetic control estimation.
result Syntax outperforms conventional trial designs in identifying beneficial subpopulations.
New loss function restores importance weighting in overparameterized models.
problem Restoring importance weighting in overparameterized neural networks.
method Introduced polynomially-tailed losses to restore effects of importance weighting.
result Polynomially-tailed losses improve performance in correcting distribution shift.
Boosts test performance with importance sampling for subpopulation issues.
problem Limited ERM performance on data with spurious correlation or subpopulation.
method Importance sampling as a tool to solve subpopulation problems.
result Achieves state-of-the-art performance on benchmark datasets.
Proposes Population Difference Criterion for visually observed subpopulation differences.
problem Statistical significance of visually observed subpopulation differences in high-dimensional and high-signal contexts.
method Balanced permutation approach and bootstrap confidence interval for quantifying uncertainty.
result Balanced permutation approach is more powerful in high-signal contexts.
New method recovers predictions from unobservable source subpopulation in binary classification.
problem Challenging binary classification with unobservable subpopulation in source domain.
method Distribution matching method to estimate subpopulation proportions, rigorous derivation of prediction models.
result Our method outperforms naive benchmarks in synthetic and real-world datasets.
BAICS identifies best arm with fairness constraints on subpopulations.
problem Identify the best arm while ensuring fairness across subpopulations.
method Formulated and solved BAICS problem, analyzed complexity, designed algorithm.
result Algorithm's sample complexity matches theoretical lower bound.
Optimal strategy for A/B/n testing with control in stratified populations.
problem Discovering which arms have higher weighted expectation than the control in stratified populations.
method Sequentially choosing one arm per time step based on weighted subpopulation means.
result The strategy is asymptotically optimal, ensuring correct answers with high probability.
MANO normalizes logits to estimate test accuracy without labels.
problem Estimating test accuracy of OOD samples without labels.
method Applies Lp norm to normalized logits. result Achieves state-of-the-art performance across various architectures.
Study adaptive clinical trial methods for identifying patient subpopulations with treatment benefit.
problem Adaptive identification of patient subpopulations with treatment benefit in clinical trials.
method Proposes AdaGGI and AdaGCPI meta-algorithms for subpopulation construction.
result Empirical investigation of AdaGGI and AdaGCPI performance across various simulation scenarios.
Subpopulation attacks poison data to misclassify naturally distributed points.
problem Improving accuracy of machine learning predictions through adversarial data modification.
method Introducing a novel subpopulation attack framework, using influence functions and gradient optimization.
result Subpopulation attacks are effective and stealthy, making them difficult to defend against.
In the recent literature on estimating heterogeneous treatment effects, each proposed method makes its own set of restrictive assumptions about the intervention's effects and which subpopulations to explicitly estimate. Moreover, the majority of the literature provides no mechanism to identify which subpopulations are …
Study evaluates approaches to improve worst-case model performance across patient subpopulations.
problem Improving model accuracy for specific patient subpopulations.
method Comparison of distributionally robust optimization (DRO) and standard learning procedures.
result Standard learning procedures generally outperform DRO approaches for improving model performance across subpopulations.
We consider the problem in regression analysis of identifying subpopulations that exhibit different patterns of response, where each subpopulation requires a different underlying model. Unlike statistical cohorts, these subpopulations are not known a priori; thus, we refer to them as cadres. When the cadres and their a…
New method reduces variance in subpopulation model performance estimates.
problem High variance in subpopulation performance metrics for small groups.
method Using an evaluation model to form model-based metric (MBM) estimates.
result MBMs produce more accurate and lower variance estimates for small subpopulations.
Paper proposes a novel MTL framework for personalized modeling of diverse individuals.
problem Personalized modeling of heterogeneous subpopulations with high-dimensional data.
method Low-rank tensor decomposition for multi-task learning.
result Superior performance compared to benchmarks in diverse subpopulation scenarios.
New split rules improve subpopulation targeting in policy-making.
problem Improving binary classification for subpopulation targeting in policy-making.
method MDFS, PFS, wEFS for maximizing distance and penalizing final splits.
result Proposed methods target more vulnerable subpopulations than classic CART/KD-CART.
Conventional survival analysis approaches estimate risk scores or individualized time-to-event distributions conditioned on covariates. In practice, there is often great population-level phenotypic heterogeneity, resulting from (unknown) subpopulations with diverse risk profiles or survival distributions. As a result, …
Unsupervised recalibration (URC) is a general way to improve the accuracy of an already trained probabilistic classification or regression model upon encountering new data while deployed in the field. URC does not require any ground truth associated with the new field data. URC merely observes the model's predictions a…
ATC predicts target domain accuracy using only labeled and unlabeled data.
problem Predicting out-of-distribution performance with limited labeled data.
method Average Thresholded Confidence (ATC) method that learns a threshold on model confidence.
result ATC outperforms previous methods across various types of distribution shifts and datasets.
Efficiently audits model fairness with continuous monitoring and flexible data collection.
problem Continuous monitoring and flexible data collection for fairness auditing.
method Sequential, anytime-valid inference and game-theoretic statistics.
result Demonstrated efficacy on three fairness datasets.
S2M optimizes mining for diverse data subpopulations.
problem Scalability and uniformity in training sets with many labels and diverse data.
method Doubly-stochastic mining (S2M) computes per-example and minibatch losses on hardest labels/examples.
result S2M ensures good performance across all data subpopulations.
The study assesses ML model robustness under worst-case subpopulations.
problem ML model performance degradation under non-training population.
method Two-stage estimation procedure for evaluating worst-case robustness over subpopulations.
result The method certifies model robustness and prevents unreliable deployments.
Data augmentation methods improve worst-case model performance.
problem Ensuring fair predictions across subpopulations in large models.
method Linear last layer retraining with data augmentation techniques.
result Optimal worst-group accuracy achieved for Gaussian latent representation distribution.
Models predict probabilities of causation from limited data.
problem Estimating probabilities of causation requires unreliable or impractical experimental and observational data.
method Proposed Exact-MLP and Mask-MLP models trained on reliable subpopulations.
result Models achieve average MAEs of roughly 0.03, reducing MAE by 80%.
Develops CPL for optimal prediction set length and validity.
problem Balancing conditional validity and length efficiency in conformal prediction.
method Conformal Prediction with Length-Optimization (CPL).
result Achieves optimal prediction set length while maintaining conditional validity.
Study causal effects on humans in mixed human-AI systems with unobserved unit types.
problem Estimating causal effects on humans in systems with unobserved unit types and interaction networks.
method Assumed human-AI prior, causal message passing (CMP) framework, subpopulation analysis.
result Consistently recover human-specific causal effects using subpopulations with varying expected human composition and treatment exposure.
Imputation-Powered Inference improves subpopulation efficiency in missing data settings.
problem Complex missing data patterns challenge standard inference methods.
method Imputation-Powered Inference (IPI) combines blackbox imputation with bias correction.
result IPI provides valid and efficient M-estimation under MCAR blockwise missingness.
RAD improves robustness to domain annotation noise without explicit domain annotations.
problem Robustness to domain annotation noise in training data.
method Regularized Annotation of Domains (RAD) for last layer retraining.
result RAD outperforms state-of-the-art methods even with 5% noise in training data.
Two new models forecast multiple subpopulations' mortality, outperforming existing methods.
problem Modeling mortality patterns and trajectories in related populations.
method Adapted multivariate functional principal component analysis techniques.
result The second proposed model outperforms existing methods in forecast accuracy.
Tumors often contain multiple subpopulations of cancerous cells defined by distinct somatic mutations. We describe a new method, PhyloWGS, that can be applied to WGS data from one or more tumor samples to reconstruct complete genotypes of these subpopulations based on variant allele frequencies (VAFs) of point mutation…
We consider the problem in precision health of grouping people into subpopulations based on their degree of vulnerability to a risk factor. These subpopulations cannot be discovered with traditional clustering techniques because their quality is evaluated with a supervised metric: the ease of modeling a response variab…
Develops methods to analyze feature-outcome associations in subpopulations.
problem Challenges in understanding feature-outcome associations in high-dimensional data.
method Geometric decomposition framework using gradient flow and co-monotonicity decomposition.
result Identifies context-dependent patterns and improves statistical power and interpretability.
One primary task of population health analysis is the identification of risk factors that, for some subpopulation, have a significant association with some health condition. Examples include finding lifestyle factors associated with chronic diseases and finding genetic mutations associated with diseases in precision he…
Study evaluates information leakage in Polymarket markets, finding limited applicability and resolution ambiguity.
problem Limited applicability of ILS-dl framework across Polymarket markets.
method Scaling from single-case to population-scale evaluation using ILS-dl framework.
result Only 0.7% of candidate markets yield computable ILS-dl values, and resolution semantics are the main obstacle.
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.
Study identifies and estimates treatment effect heterogeneity within principal stratification subpopulations.
problem Causal inference with intermediate outcomes and treatment effect heterogeneity.
method Proposes a novel doubly cross-fit doubly robust machine learner to efficiently learn conditional principal causal effects under principal ignorability.
result Demonstrates informative patterns of treatment effect heterogeneity within the always-survivor subpopulation in an acute lung injury trial.