arXiv Machine Learning

Sufficiently Reduced Distributional Regression

Sufficiently Reduced Distributional Regression (SRDR) is a generative approach that merges conditional distribution estimation with nonlinear sufficient dimension reduction (SDR). By framing SDR as a risk minimization problem using strictly proper scoring rules, SRDR jointly learns a dimension reduction map and a generative prediction model through minimization of the energy score, which can be estimated via sampling. The method extends to multi‑environment data and classification, and theoretical results show convergence of estimated conditional distributions in energy distance, implying asymptotic sufficiency. In experiments on CT slice localization, superconductivity data, and digit classification, SRDR recovers low‑dimensional sufficient structure and matches or surpasses state‑of‑the‑art nonlinear SDR methods in representation quality and predictive performance.

arXiv Machine Learning
Oct 28

Active Learning of Conditional Generative Models via the Transport Neural Tangent Kernel

The paper introduces the transport Neural Tangent Kernel (tNTK) to enable active learning for conditional generative models that produce distributional outputs. By decomposing the Wasserstein test risk into mean and shape components, the authors show that mean‑based active learning misses distributional shape, and propose tNTK as a tractable kernel that captures sensitivity of the predicted distribution to model parameters. Using tNTK as a Gaussian process covariance, they greedily select inputs that minimize posterior variance, and demonstrate that this distribution‑aware strategy outperforms mean‑based NTK baselines across several generative model types and datasets.

By Jayoung Ryu, Kyunghyun Cho, Romain Lopez
arXiv Machine Learning
Aug 24

Conditional-Independence-Regularized Distributional Autoencoders for Mixed-Type Data

The paper introduces Conditional-Independence-Regularized Distributional Autoencoders, a framework for learning low-dimensional representations of mixed-type data that includes both numerical and categorical variables. It uses an energy-score objective for numerical variables, a likelihood objective for categorical variables, and an auxiliary conditional independence regularization term to capture dependencies between variable types. The authors provide theoretical analysis and demonstrate that the method improves categorical distribution recovery, achieves competitive overall conditional distribution recovery, and preserves mixed-type dependence structure on synthetic and real-world datasets.

By Siyuan Tang, Gongjun Xu, Ji Zhu
arXiv Machine Learning
Oct 9

Marformer: A Transformer for Predicting Missing Data Distributions

The Marformer is a Transformer model designed to predict conditional marginal distributions of missing variables given observed data, enabling computation of Bayes risk and Value of Information. It learns a hidden-vector representation for each variable’s distribution and refines these through attention, similar to BERT’s masked language modeling but for probabilistic inference. Evaluations on synthetic Bayesian networks, discretized Gaussians, structured annotation data, and a real annotation dataset show that the Marformer matches or surpasses classical missing-data methods while being faster and not requiring explicit joint distribution modeling.

By Prabhav Singh, Xiheng Tom Wang, Haojun Shi, Jason Eisner
arXiv Machine Learning
Aug 26

Generalization, memorization, and overfitting for diffusion models trained in the lazy high-dimensional regime

The paper investigates diffusion models trained in a lazy high‑dimensional regime, extending benign overfitting theory to generative settings. By analyzing denoising score matching in a vector‑valued RKHS with an inner‑product kernel, the authors derive exact risk trajectories under gradient flow when the number of samples scales proportionally with dimensionality. These trajectories reveal three distinct phases—spectral generalization, noise‑dominated interpolation, and empirical Bayes memorization—whose interplay shapes the distribution of generated samples.

By Hugo Latourelle-Vigeant, Sinho Chewi, Aram-Alexandre Pooladian, John Sous, Theodor Misiakiewicz
arXiv Machine Learning
Sep 1

Prediction-Powered Conditional Inference

arXiv:2603.05575v2 Announce Type: replace-cross Abstract: We study prediction-powered conditional inference in the setting where labeled data are scarce, unlabeled covariates are abundant, and a blac...

By Yang Sui, Jin Zhou, Hua Zhou, Xiaowu Dai