A causal process model for algorithmic recourse introduces post-recourse stability conditions and copula-based methods to infer intervention effects from observational or paired data, with a distribution-free fallback when the model is rejected.
hub
Machine learning , volume=
26 Pith papers cite this work. Polarity classification is still indexing.
26
Pith papers citing it
hub tools
citation-role summary
background 2
method 1
citation-polarity summary
years
2026 26representative citing papers
Randomized calibrated expert aggregation that Blackwell-refines a target is polynomial-time solvable; deterministic aggregation is NP-hard and has no multiplicative PTAS for proper losses.
A methodological framework for separable effects analysis that distinguishes four-arm and two-arm designs, with EIF-based estimation and falsification tests.
PAIR-CI restores calibration to conditional independence testing under missing data by using paired permutations that force imputation error to cancel in the loss difference, together with a consistent variance estimator that unifies cross-validation and imputation uncertainty.
A conditional adaptive perturbation approach enables valid in-sample inference for machine learning-identified subgroups with nonregular boundaries via triple robustness.
Temperature scaling of density-matrix eigenvalues from LLM semantic embeddings optimizes proper-score calibration and corrects systematic overconfidence so entropy equals risk.
MAPLE estimates conditional class probabilities by local averaging over Mapper-graph neighborhoods with data-driven cover selection and proves consistency under regularity conditions.
Di-COT is an unsupervised contrastive method that stochastically partitions time-series windows into overlapping sub-blocks to learn representations without augmentation, reporting SOTA results on classification and transfer tasks across multiple benchmarks while cutting training time.
Introduces nLoI and four complementary divergence measures with within/between-node decomposition and unified permutation testing to evaluate surrogate reconstruction quality for Explainable Ensemble Trees.
PerturbedVAE disentangles perturbation-specific signals from invariant gene expression structure to recover causal representations and improve out-of-distribution prediction in single-cell perturbation modeling.
An augmented kernel ridge regression estimator separates linear and nonlinear components to achieve sharp oracle inequalities and minimax optimal prediction risk under general kernels.
TAP couples a learner-conditioned policy with diffusion inpainting to generate and selectively inject high-utility tabular augmentations, yielding up to 15.6 pp accuracy gains and 32% RMSE reduction on seven datasets under severe scarcity.
A calibration procedure yields a weighted transported average treatment effect with asymptotically valid and efficient inference when experimental data grows slower than observational data, even without positivity or correct OLS specification.
DKPS-based methods predict new model benchmark scores using cached responses, matching baseline mean absolute error with substantially fewer queries and an offline query selection approach.
Dual-Glob applies supervised contrastive learning to classify fine-grained pitch accent patterns from F0 contours in Seoul Korean, achieving 77.75% accuracy and 51.54% F1 on a new dataset of 10,093 manually annotated accentual phrases.
GeoFlow improves OD flow prediction and generation by augmenting area representations with geospatial attributes and using a geometric-intrinsic fusion encoder with axial-global attention decoder.
ERPPO adds a DSA-based ambiguity estimator to MAPPO and switches between L1 and L2 entropy regularization to improve exploration and stability in non-stationary multi-dimensional observations.
Graph-augmented LLMs using a political knowledge graph improve ideology prediction accuracy for Swiss MPs by incorporating relational data beyond text alone.
Adaptive MSD-Splitting improves C4.5 and Random Forest performance on skewed data by adjusting standard deviation multipliers for discretization while retaining linear time complexity.
TabGRAA applies group-relative advantage alignment in an iterative reward-guided post-training loop to improve tabular language model generators on fidelity, utility, and privacy trade-offs across five benchmarks.
On five tabular security datasets at 10% labels, tuning only the classifier with Bayesian optimization recovers a median 86% of the gains from full joint SSL-classifier optimization.
Frontal and fronto-central EEG regions show the most consistent predictive utility for cognitive workload in subject-independent settings, outperforming full-scalp baselines by 15-20% in relative rank across datasets.
The work introduces WaLeF/FIDLAr for flood forecasting, CoDiCast for probabilistic weather, and Hypercube-RAG for explainable environmental QA, claiming superior accuracy, efficiency, and interpretability over baselines.
The paper introduces a question-driven framework and set of statistical methods for exploratory assessment of regional treatment effect heterogeneity in multi-regional clinical trials, evaluated via simulations under no-heterogeneity and modifier-driven scenarios.
citing papers explorer
-
Causal Algorithmic Recourse: Foundations and Methods
A causal process model for algorithmic recourse introduces post-recourse stability conditions and copula-based methods to infer intervention effects from observational or paired data, with a distribution-free fallback when the model is rejected.
-
Algorithmic Expert Aggregation
Randomized calibrated expert aggregation that Blackwell-refines a target is polynomial-time solvable; deterministic aggregation is NP-hard and has no multiplicative PTAS for proper losses.
-
Separable Effects in Four-Arm and Two-Arm Designs
A methodological framework for separable effects analysis that distinguishes four-arm and two-arm designs, with EIF-based estimation and falsification tests.
-
PAIR-CI: Calibrated Conditional Independence Testing for Causal Discovery with Incomplete Data
PAIR-CI restores calibration to conditional independence testing under missing data by using paired permutations that force imputation error to cancel in the loss difference, together with a consistent variance estimator that unifies cross-validation and imputation uncertainty.
-
In-Sample Evaluation of Subgroups Identified by Generic Machine Learning
A conditional adaptive perturbation approach enables valid in-sample inference for machine learning-identified subgroups with nonregular boundaries via triple robustness.
-
Eigenvalue Calibration for Semantic Embeddings of Large Language Models
Temperature scaling of density-matrix eigenvalues from LLM semantic embeddings optimizes proper-score calibration and corrects systematic overconfidence so entropy equals risk.
-
MAPLE: Mapper Based Localized Prediction with Data Driven Cover Selection for High dimensional Data
MAPLE estimates conditional class probabilities by local averaging over Mapper-graph neighborhoods with data-driven cover selection and proves consistency under regularity conditions.
-
Divide and Contrast: Learning Robust Temporal Features without Augmentation
Di-COT is an unsupervised contrastive method that stochastically partitions time-series windows into overlapping sub-blocks to learn representations without augmentation, reporting SOTA results on classification and transfer tasks across multiple benchmarks while cutting training time.
-
A Family of Divergence Measures for Evaluating the Reconstruction Quality of Explainable Ensemble Trees
Introduces nLoI and four complementary divergence measures with within/between-node decomposition and unified permutation testing to evaluate surrogate reconstruction quality for Explainable Ensemble Trees.
-
What Makes a Representation Good for Single-Cell Perturbation Prediction?
PerturbedVAE disentangles perturbation-specific signals from invariant gene expression structure to recover causal representations and improve out-of-distribution prediction in single-cell perturbation modeling.
-
Adaptive Kernel Ridge Regression with Linear Structure: Sharp Oracle Inequalities and Minimax Optimality
An augmented kernel ridge regression estimator separates linear and nonlinear components to achieve sharp oracle inequalities and minimax optimal prediction risk under general kernels.
-
Active Tabular Augmentation via Policy-Guided Diffusion Inpainting
TAP couples a learner-conditioned policy with diffusion inpainting to generate and selectively inject high-utility tabular augmentations, yielding up to 15.6 pp accuracy gains and 32% RMSE reduction on seven datasets under severe scarcity.
-
Transporting treatment effects by calibrating large-scale observational outcomes
A calibration procedure yields a weighted transported average treatment effect with asymptotically valid and efficient inference when experimental data grows slower than observational data, even without positivity or correct OLS specification.
-
Query-efficient model evaluation using cached responses
DKPS-based methods predict new model benchmark scores using cached responses, matching baseline mean absolute error with substantially fewer queries and an offline query selection approach.
-
Deep Supervised Contrastive Learning of Pitch Contours for Robust Pitch Accent Classification in Seoul Korean
Dual-Glob applies supervised contrastive learning to classify fine-grained pitch accent patterns from F0 contours in Seoul Korean, achieving 77.75% accuracy and 51.54% F1 on a new dataset of 10,093 manually annotated accentual phrases.
-
GeoFlow: Geo-Aware Modeling of Inter-Area Relationships in Origin-Destination Flow Prediction and Generation
GeoFlow improves OD flow prediction and generation by augmenting area representations with geospatial attributes and using a geometric-intrinsic fusion encoder with axial-global attention decoder.
-
ERPPO: Entropy Regularization-based Proximal Policy Optimization
ERPPO adds a DSA-based ambiguity estimator to MAPPO and switches between L1 and L2 entropy regularization to improve exploration and stability in non-stationary multi-dimensional observations.
-
Graph-Augmented LLMs for Swiss MP Ideology Prediction
Graph-augmented LLMs using a political knowledge graph improve ideology prediction accuracy for Swiss MPs by incorporating relational data beyond text alone.
-
Adaptive MSD-Splitting: Enhancing C4.5 and Random Forests for Skewed Continuous Attributes
Adaptive MSD-Splitting improves C4.5 and Random Forest performance on skewed data by adjusting standard deviation multipliers for discretization while retaining linear time complexity.
-
Self-Improving Tabular Language Models via Iterative Reward-Guided Post-Training
TabGRAA applies group-relative advantage alignment in an iterative reward-guided post-training loop to improve tabular language model generators on fidelity, utility, and privacy trade-offs across five benchmarks.
-
SemiScope: Disentangling Classifier Tuning and Joint Optimization in Semi-Supervised Security Classification
On five tabular security datasets at 10% labels, tuning only the classifier with Bayesian optimization recovers a median 86% of the gains from full joint SSL-classifier optimization.
-
Assessing Region-Level EEG Contributions to Cognitive Workload Prediction
Frontal and fronto-central EEG regions show the most consistent predictive utility for cognitive workload in subject-independent settings, outperforming full-scalp baselines by 15-20% in relative rank across datasets.
-
Accurate, Efficient, and Explainable Deep Learning Approaches for Environmental Science Problems
The work introduces WaLeF/FIDLAr for flood forecasting, CoDiCast for probabilistic weather, and Hypercube-RAG for explainable environmental QA, claiming superior accuracy, efficiency, and interpretability over baselines.
-
A Workflow for Evaluating Regional Treatment Effect Heterogeneity in Multi-Regional Clinical Trials
The paper introduces a question-driven framework and set of statistical methods for exploratory assessment of regional treatment effect heterogeneity in multi-regional clinical trials, evaluated via simulations under no-heterogeneity and modifier-driven scenarios.
-
Generative Augmentation of Imbalanced Flight Records for Flight Diversion Prediction: A Multi-objective Optimisation Framework
A domain-tuned multi-objective search over generative-model hyperparameters improves synthetic flight-diversion data quality and boosts diversion prediction versus real-data-only training.
- Causal Stability Selection