Bayesian algorithm detects changes in fluctuating baselines.
problem Detecting change points in time series with a shifting baseline.
method Extended Bayesian online change point detection (BOCPD) algorithm.
result The extended algorithm can detect changes in fluctuating baselines.
Bayesian Scattering offers a simple baseline for image data uncertainty.
problem Lack of interpretable, mathematically grounded uncertainty quantification methods for image data.
method Coupling wavelet scattering transform with a simple probabilistic head.
result Bayesian Scattering provides sensible uncertainty estimates under distribution shifts.
Novel neural models have been proposed in recent years for learning under domain shift. Most models, however, only evaluate on a single task, on proprietary datasets, or compare to weak baselines, which makes comparison of models difficult. In this paper, we re-evaluate classic general-purpose bootstrapping approaches …
Label shift refers to the phenomenon where the prior class probability p(y) changes between the training and test distributions, while the conditional probability p(x|y) stays fixed. Label shift arises in settings like medical diagnosis, where a classifier trained to predict disease given symptoms must be adapted to sc…
Shifts dataset evaluates uncertainty in real-world tasks across modalities.
problem Lack of standard datasets for evaluating uncertainty estimation and robustness to distributional shift.
method Proposes Shifts Dataset for evaluation of uncertainty estimates and robustness to distributional shift across tabular, audio, text, and sensor data.
result Baseline results for tabular weather prediction, machine translation, and SDC vehicle motion prediction.
New method reduces label and data shifts between domains using optimal transport.
problem Label shift between source and target domains in domain adaptation.
method Developed theory and LDROT method to mitigate label and data shifts.
result Theoretical and experimental validation of LDROT's effectiveness.
This work evaluates uncertainty in deep Gaussian processes.
problem Uncertainty quantification in deep Gaussian processes.
method Hierarchical deep Gaussian processes (DGPs) and Deep Sigma Point Processes (DSPPs) evaluated on regression and classification tasks.
result DSPPs provide strong in-distribution calibration but are less robust under distribution shift compared to ensembles.
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.
Extends Shifts dataset for MS lesion segmentation and marine vessel power estimation.
problem Distributional shift in training and deployment data for ML models.
method Develops new datasets for high-risk industrial applications.
result Demonstrates robustness and uncertainty estimation in new industrial tasks.
Proposes SGShift to identify shifted features causing model performance degradation under concept shift.
problem Concept shift leading to miscalibration in ML models across domains.
method SGShift method for identifying sparse set of shifted features using feature selection and statistical tools.
result SGShift identifies shifted features more accurately than baseline methods, requires few samples in the shifted domain, and is robust to complex cases.
FDN improves probabilistic regressors' adaptability to distribution shifts.
problem Overconfidence in modern probabilistic regressors under distribution shift.
method FDN uses input-conditioned distributions over network weights, trained with a Monte Carlo beta-ELBO objective.
result FDN produces predictive mixtures whose dispersion adapts to the input, providing shift-aware uncertainty.
CP2 uses geometric information to improve conformal prediction robustness.
problem CP fails under geometric data shifts, losing coverage guarantees.
method Integrates geometric pose information into CP via canonicalization.
result Integrating geometric information with CP ensures robustness under geometric shifts.
IUPM monitors machine learning models under gradual shifts using optimal transport and active labeling.
problem Gradual distribution shifts lead to unnoticed accuracy declines in machine learning models.
method Incremental Uncertainty-aware Performance Monitoring (IUPM) using optimal transport and active labeling.
result IUPM outperforms existing baselines in gradual shift scenarios and guides label acquisition more effectively.
Drift-Resilient TabPFN learns to adapt to changing data distributions.
problem Real-world data often shifts over time, degrading model performance.
method In-Context Learning with a Prior-Data Fitted Network, using structural causal models.
result Significant performance improvements across various datasets.
A feature-weighted mean shift algorithm improves clustering in high-dimensional data.
problem Clustering high-dimensional data with traditional mean shift algorithms.
method Feature-weighted mean shift algorithm.
result The algorithm outperforms conventional mean shift and preserves computational simplicity.
HySRL improves RL sample efficiency with shifted-dynamics data.
problem Leveraging historical data with shifted dynamics to improve sample efficiency in RL.
method HySRL, a hybrid transfer RL algorithm that uses prior information on dynamics shift to achieve better sample complexity.
result HySRL achieves problem-dependent sample complexity and outperforms pure online RL.
CTRF combines logged data and randomized experiments for robust prediction.
problem Robust prediction models to handle distributional shifts between training and testing data.
method CTRF uses existing training data and a small amount of randomized experiment data to train a robust model.
result CTRF produces robust predictions and outperforms baseline methods in the presence of feature shifts.
Paper improves prediction sets for distribution shifts without labels.
problem Improving prediction sets effectiveness in the presence of distribution shifts.
method Develops ECP and EACP methods to adjust score function based on model uncertainty.
result Consistent improvement over existing baselines and nearly matches fully supervised methods.
RIA method improves OoD generalization for covariate shift.
problem Improving out-of-distribution generalization under covariate shift.
method Adversarial label invariant graph data augmentations for OoD generalization.
result RIA method achieves high accuracy compared to OoD baselines.
RLSbench benchmarks domain adaptation under label proportion shifts, revealing widespread failures and proposing a two-step meta-algorithm.
problem Domain adaptation under label proportion shifts is poorly understood and inconsistent across methods.
method RLSbench introduces a large-scale benchmark with 500 distribution shift pairs. It proposes a two-step meta-algorithm to improve domain adaptation methods under label proportion shifts.
result The two-step meta-algorithm improves domain adaptation methods by 2-10% accuracy points under large label proportion shifts.
Unified framework for distribution shift estimation, explanation, and improvement.
problem Estimating, explaining, and improving model performance on target domains with distribution shift.
method Entropic Projection Alignment (EPA) aligns source and target distributions by matching moments and minimizing KL divergence.
result EPA consistently outperforms state-of-the-art baselines while offering computational efficiency.
FIT evaluates time series model feature importance quantifying distributional shift.
problem Lack of explanations for time series models in high-stakes applications.
method FIT framework quantifies feature importance based on distributional shift using KL-divergence.
result FIT identifies important time points and observations superiorly compared to baselines.
New algorithm identifies best arm in semiparametric bandits with near optimal efficiency.
problem Fixed-confidence Best Arm Identification in semiparametric bandits with unknown baseline shift.
method Phase-elimination algorithm based on orthogonalized regression design.
result Nearly optimal high-probability sample-complexity upper bound established.
MAGDiff detects data shifts in neural networks without retraining.
problem Neural networks' sensitivity to data distribution shifts.
method Extracts MAGDiff representations from neural networks to detect shifts.
result MAGDiff representations improve data set shift detection.
Current methods for covariate-shift adaptation use unlabelled data to compute importance weights or domain-invariant features, while the final model is trained on labelled data only. Here, we consider a particular case of covariate shift which allows us also to learn from unlabelled data, that is, combining adaptation …
Adaptive monitoring for AI systems detects and diagnoses shifts in data distribution.
problem Continuous monitoring of AI systems to detect and address unsafe behavior.
method Weighted-conformal martingales (WCTMs) for online monitoring of AI systems.
result Improved performance over state-of-the-art baselines on real-world datasets.
The paper addresses missing data imputation issues by correcting for distribution shift.
problem Missing data imputation and the resulting distribution shift between observed and full data.
method Formulates imputation as a risk minimization problem and proposes a novel algorithm to correct for distribution shift.
result The proposed algorithm consistently improves imputation accuracy, reducing RMSE and Wasserstein distance by 3% and 7%, respectively.
Proposes a method to correct for covariate shift in meta-analysis of randomized trials.
problem Invalidation of standard IPD meta-analysis due to covariate shift across studies.
method Placebo-anchored transport framework that treats source-trial outcomes as proxy signals and target-trial placebo outcomes as gold labels.
result Yields target-identified effect estimates in connected targets and a principled screen--then--transport procedure in disconnected targets.
Evaluating prediction models under covariate shift and selective labels
problem Model performance evaluation under distribution shift and selection bias
method Double machine learning
result Accurate estimation of target risk
Optimal selective classification using likelihood ratios improves model reliability.
problem Enhancing predictive model reliability by allowing uncertain predictions.
method Neyman--Pearson lemma applied to likelihood ratios for optimal selection.
result Neyman--Pearson-informed methods outperform existing baselines under covariate shifts.
New method certifies risks of LLM outputs, improving accuracy and reliability.
problem Uncertain and incorrect outputs from large language models.
method Information-lift certificates using PAC-Bayes bounds and skeleton design.
result Achieves 77.0% coverage at 2% risk, outperforming baselines.
A new method uses nearest neighbors for importance weighting.
problem Data covariate shift problems in machine learning.
method Nearest neighbor classification scheme for determining importance weights.
result Demonstrated effectiveness through comparative experiments on various classification tasks.
Enhanced regime shifts detection using unstructured text and financial data.
problem Detecting regime shifts in financial markets is challenging due to noisy and multicollinear data.
method Combines LLM reasoning on unstructured text and statistical validation on financial time series.
result Framework achieves F1 score of 0.82, outperforming pure data-driven methods.
Undersampling often outperforms other methods in nonparametric classification.
problem Distribution shift challenges in nonparametric binary classification.
method Proved undersampling is minimax optimal in worst-case scenarios.
result Undersampling is a robustness intervention with theoretical guarantees.
Improved OOD detection across various shifts using multi-encoder fusion of RDMs.
problem Out-of-distribution detection across multiple types of distribution shifts.
method Statistical identification of encoder sensitivity, EncMin2L fusion, and Tippett minimum combination.
result Achieves AUROC ≥ 0.94 across four shift types, outperforming state-of-the-art detectors.
Proposes BSSP to stabilize predictions in biased data.
problem Distribution shift between training and test data causes prediction instability.
method Balance-subsampled stable prediction (BSSP) algorithm based on fractional factorial design.
result Significantly improves prediction stability across unknown test data.
Causal Imitation Learning handles noisy measurements and distribution shifts.
problem Learning from noisy state observations and distributional shifts.
method Causal inference framework and adversarial RKHS learning.
result Improved robustness to distribution shifts compared to standard methods.
It has been shown that instead of learning actual object features, deep networks tend to exploit non-robust (spurious) discriminative features that are shared between training and test sets. Therefore, while they achieve state of the art performance on such test sets, they achieve poor generalization on out of distribu…
FLUXtrapolation benchmarks machine learning for extrapolating ecosystem fluxes under distribution shifts.
problem Machine learning challenges in extrapolating ecosystem fluxes under distribution shifts.
method Defined temporal, spatial, and temperature-based extrapolation scenarios; evaluated performance across domains, temporal aggregations, and tail errors.
result Baselines perform similarly under median hourly RMSE but differ under tail-focused and multi-scale evaluations.
Algorithm constructs prediction sets with PAC guarantees in label shift settings.
problem Reliable uncertainty quantification in the face of distribution shift.
method Estimates predicted probabilities and confusion matrix, then propagates uncertainty through Gaussian elimination to compute confidence intervals and construct prediction sets.
result Satisfies PAC guarantees and produces smaller, more informative prediction sets.
The paper tackles matching a desired mean in causal systems through shift interventions.
problem Matching a desired mean in causal systems.
method Defining Markov equivalence classes, proposing active learning strategies, deriving lower bounds.
result Proposed active learning strategies require fewer interventions than previous approaches, especially for certain graph classes.
Bayesian framework estimates label shift for improved classifier performance.
problem Label shift in supervised learning leading to degraded classifier performance.
method Bayesian framework with dynamic Dirichlet priors and online EM algorithms.
result Significant improvements in classifier accuracy over state-of-the-art methods.
Adapts example weights to optimize black-box metrics.
problem Optimizing metrics defined by black-box functions.
method Adaptive example weighting and iterative post-shifting.
result Improves classification performance compared to baselines.
In safety-critical applications of machine learning, it is often important to abstain from making predictions on low confidence examples. Standard abstention methods tend to be focused on optimizing top-k accuracy, but in many applications, accuracy is not the metric of interest. Further, label shift (a shift in class …
RAVEN improves weak-to-strong generalization under distribution shifts.
problem Weak models fail to supervise strong models effectively under distribution shifts.
method RAVEN dynamically learns optimal combinations of weak models and strong model parameters.
result RAVEN outperforms existing methods by over 30% on out-of-distribution tasks.
CATS adapts multivariate time series models by addressing correlation shift.
problem Correlation differences across domains in multivariate time series data.
method CATS introduces correlation shift to measure domain differences, and uses a graph attention module and temporal convolution to align target correlations with source correlations.
result CATS increases over 10% average accuracy compared to vanilla Transformer-based models with minimal additional parameters.
Anchor-TS uses median anchoring to improve online decision-making from offline data with distribution shift.
problem Improving online decision-making from offline data with distribution shift.
method Sample-Mean Anchored Thompson Sampling (Anchor-TS) with median anchoring.
result Anchor-TS safely leverages offline data to accelerate online learning and reduces regret.
Method learns diverse models to resist spurious correlations.
problem Models learn spurious correlations in i.i.d. training data.
method Multiple models with minimized conditional total correlation.
result Framework significantly outperforms baseline on distribution shift.