Multi-source transfer learning incurs an intrinsic adaptation cost that can exceed one, with phase transitions separating regimes where bias-agnostic estimators match oracle performance from those where they cannot.
Graphical Models for Processing Missing Data
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
Introduces bridge score as balancing score for mediator ignorability, derives sharp pointwise bounds on mediator-outcome confounding via two latent parameters, and provides benchmark and residual-budget calibration for sensitivity analysis.
Introduces decision-aware proximal bridge learning using a weighted loss and regret bound to enhance optimal treatment selection in settings with hidden confounding.
LGR samples balanced treatment assignments in high-dimensional experiments via continuous relaxation and SGLD, retaining valid inference through randomization tests while being orders of magnitude faster than prior methods.
Derives asymptotic efficiency bounds for a broad class of sequential experimental designs showing no further first-order asymptotic efficiency gains are possible for ATE estimation beyond the Hahn (1998) bound achieved with optimized propensity scores.
A recursive Riesz representer-based targeted minimum loss estimation procedure unifies asymptotically efficient estimation of causal estimands such as time-varying treatment effects and mediation effects.
A novel bias-reduced online covariance estimator for SGD achieves convergence rate n to the power (α-1)/2 times square root of log n without second-order derivatives.
Geometric tempering yields exponential convergence bounds for both Wasserstein and Fisher-Rao flows but produces no speedup in the Fisher-Rao metric, with new adaptive schedules derived from the tempered dynamics.
Introduces self-separated and self-connected missingness models for mediator and outcome missingness in mediation analysis, enabling identification via conditional independences or shadow variables and extending shadow variable theory.
Local L1-regularized robust estimators plus server-side robust aggregation achieve near-optimal rates for high-dimensional sparse learning under Byzantine attacks.
A mirror descent algorithm computes exact Wasserstein barycenters for mixed discrete and continuous input measures with convergence guarantees.
A review and visual synthesis of links between network models and latent variable approaches in psychometrics, proposing extensions to the methodological toolbox via cross-domain methods.
The crumble package provides nonparametric tools for estimating natural direct/indirect effects, randomized interventional effects, and recanting-twin effects in mediation analysis, with guidance on identification assumptions and non-binary treatments illustrated via case studies.
citing papers explorer
-
The Statistical Cost of Adaptation in Multi-Source Transfer Learning
Multi-source transfer learning incurs an intrinsic adaptation cost that can exceed one, with phase transitions separating regimes where bias-agnostic estimators match oracle performance from those where they cannot.
-
Sensitivity analysis for causal mediation: bridge score, sharp sensitivity bounds, and calibration
Introduces bridge score as balancing score for mediator ignorability, derives sharp pointwise bounds on mediator-outcome confounding via two latent parameters, and provides benchmark and residual-budget calibration for sensitivity analysis.
-
Decision-Aware Proximal Bridge Learning for Optimal Treatment Selection
Introduces decision-aware proximal bridge learning using a weighted loss and regret bound to enhance optimal treatment selection in settings with hidden confounding.
-
Langevin-Gradient Rerandomization
LGR samples balanced treatment assignments in high-dimensional experiments via continuous relaxation and SGLD, retaining valid inference through randomization tests while being orders of magnitude faster than prior methods.
-
Asymptotic Efficiency Bounds for a Class of Experimental Designs
Derives asymptotic efficiency bounds for a broad class of sequential experimental designs showing no further first-order asymptotic efficiency gains are possible for ATE estimation beyond the Hahn (1998) bound achieved with optimized propensity scores.
-
A Riesz Representer Perspective on Targeted Learning
A recursive Riesz representer-based targeted minimum loss estimation procedure unifies asymptotically efficient estimation of causal estimands such as time-varying treatment effects and mediation effects.
-
Refining Covariance Matrix Estimation in Stochastic Gradient Descent Through Bias Reduction
A novel bias-reduced online covariance estimator for SGD achieves convergence rate n to the power (α-1)/2 times square root of log n without second-order derivatives.
-
Properties and limitations of geometric tempering for gradient flow dynamics
Geometric tempering yields exponential convergence bounds for both Wasserstein and Fisher-Rao flows but produces no speedup in the Fisher-Rao metric, with new adaptive schedules derived from the tempered dynamics.
-
Self-separated and self-connected models for mediator and outcome missingness in mediation analysis
Introduces self-separated and self-connected missingness models for mediator and outcome missingness in mediation analysis, enabling identification via conditional independences or shadow variables and extending shadow variable theory.
-
Byzantine-Robust Distributed Sparse Learning Revisited
Local L1-regularized robust estimators plus server-side robust aggregation achieve near-optimal rates for high-dimensional sparse learning under Byzantine attacks.
-
A Unified Approach for Computing Wasserstein Barycenters of Discrete and Continuous Measures
A mirror descent algorithm computes exact Wasserstein barycenters for mixed discrete and continuous input measures with convergence guarantees.
-
Reconciling Latent Variables and Networks: Exploring and extending the Psychometric-Toolbox
A review and visual synthesis of links between network models and latent variable approaches in psychometrics, proposing extensions to the methodological toolbox via cross-domain methods.
-
crumble: A comprehensive framework for modern causal mediation analysis with intermediate confounding
The crumble package provides nonparametric tools for estimating natural direct/indirect effects, randomized interventional effects, and recanting-twin effects in mediation analysis, with guidance on identification assumptions and non-binary treatments illustrated via case studies.