arXiv:2501. 18897v4 Announce Type: replace-cross Abstract: Generative models have achieved remarkable success across a range of applications, yet their evaluation still lacks principled uncertainty quantification.
arXiv:2512. 19510v2 Announce Type: replace Abstract: Conditional independence (CI) is central to causal inference, feature selection, and graphical modeling, yet it is untestable in many settings without additional assumptions.
The paper introduces a new estimand for conditional distributional treatment effects that captures how treatments influence the entire outcome distribution, including variance and tail risks, in a covariate-dependent manner. It presents a doubly robust estimator that is minimax optimal locally and uses it to construct a test for global homogeneity of conditional potential outcome distributions. The test accommodates discrepancies beyond the maximum mean discrepancy, guarantees valid type‑1 error, is consistent against fixed alternatives, and includes a computationally efficient, permutation‑free algorithm with exact closed‑form expressions for two natural discrepancies.
arXiv:2609.23789v1 Announce Type: new Abstract: Modern conditional generative models face significant challenges when learning complex covariate dependencies. While sufficient dimension reduction (SD...
The paper introduces a Conditional Kernel Stein Discrepancy (CKSD) that extends traditional Kernel Stein Discrepancies to evaluate goodness‑of‑fit for conditional distributions. By employing an operator‑valued kernel on the covariate space, the authors lift unconditional KSDs to settings where only a conditional target model and joint samples are available, covering both smooth manifolds and discrete spaces. They prove that the CKSD statistic vanishes exactly when the conditional model matches the true conditional distribution for almost all covariates, and demonstrate its practical effectiveness through experiments on level, power, and runtime.
arXiv:2604. 23904v3 Announce Type: replace-cross Abstract: Synthetic tabular data are often evaluated by distributional similarity, privacy distance, or train-on-synthetic-test-on-real predictive performance, but these criteria do not ensure validity for causal inference.
arXiv:2608. 10096v1 Announce Type: cross Abstract: Modern data science increasingly gives rise to hypothesis-testing problems that are not naturally formulated in terms of parameters within prespecified statistical models.
The paper introduces a unified framework for measuring divergences between conditional distributions using kernel embeddings, termed conditional maximum mean discrepancy (CMMD). It defines a family of CMMD metrics—CMMD_0, CMMD_1, and CMMD_2—each based on different RKHS embeddings, and generalizes them to a level‑s CMMD with clear assumptions and operator‑based smoothing connections. Additionally, the authors present a novel doubly robust estimator for CMMD and demonstrate through experiments that the metric captures complex conditional dependencies for statistical testing.
arXiv:2606. 18011v1 Announce Type: cross Abstract: Constraint-based causal discovery relies on repeated conditional independence tests, but fast nonparametric tests often sacrifice calibration, especially when variables depend on the conditioning set through nonlinear relationships.
arXiv:2605. 01765v2 Announce Type: replace-cross Abstract: Mediation analysis has traditionally focused on outcome-level summary contrasts, such as mean effects, which may obscure substantial distributional changes induced by complex and nonlinear causal mechanisms.
arXiv:2604. 05324v2 Announce Type: replace Abstract: Statistical evaluation aims to estimate the generalization performance of a model using held-out i.
arXiv:2606. 18993v1 Announce Type: cross Abstract: Testing conditional independence is fundamental yet intrinsically difficult: without additional assumptions, Type I error control is impossible in general.