研究园简易版 · 经典书籍 · 期刊文章 · 有原文 · 完整版(可多选筛选、搜索) · 首页
2026 年的期刊文章 · 第 41 页
共 1756 篇,按发表时间排,新的在前,每页 40 篇。这一页是 2026 到 2026。
按期刊、按主题看 →
JMLR High-Dimensional Nonparametric & Semiparametric · 发表于 2026
Muxuan Liang, Yang Ning, Maureen A Smith, Ying-Qi Zhao
摘要
Penalized empirical risk minimization with a surrogate loss function is often used to learn a high-dimensional linear decision rule in classification problems. Although much of the literature focus on the generalization error, there is a lack of inference procedures for identifying the driving factors of the estimated decision rule, especially when the surrogate loss is non-differentiable. We propose a kernel-smoothed decorrelated score to construct hypothesis tests and interval estimators for a linear decision rule estimated using a piece-wise linear surrogate loss, which has a discontinuous gradient and non-regular Hessian. Specifically, we adopt kernel approximations to smooth the discontinuous gradient near discontinuity points and approximate the non-regular Hessian of the surrogate loss. In applications where additional nuisance parameters are involved, we propose a novel cross-fitted version to accommodate flexible nuisance estimates and kernel approximations. We establish the limiting distribution of the kernel-smoothed decorrelated score and its cross-fitted version in a high-dimensional setup. Simulation and real data analysis are conducted to demonstrate the validity and the superiority of the proposed method.
JMLR Meta-Analysis & Data Integration Clinical Trials · 发表于 2026
Yujia Gu, Hanzhong Liu, Wei Ma
摘要
Randomized clinical trials are the gold standard for analyzing treatment effects. However, increasing costs and ethical concerns may limit trial recruitment, resulting in insufficient sample sizes and potentially invalid inference. Incorporating external trial data with similar characteristics (treatments, diseases, biomarkers, etc.) into the analysis appears promising for addressing these issues. Transfer learning, which in our context utilizes external trials as the source domain and current trials as the target domain, may offer a viable approach. In this paper, we present a formal framework for applying transfer learning to the analysis of clinical trials, considering three key perspectives: transfer algorithm, theoretical foundation, and inference method. To cover broad types of randomized trials, we study this problem under stratified randomization, or more generally, covariate-adaptive randomization. For the algorithm, we adopt a parameter-based transfer learning approach to enhance the lasso-adjusted stratum-specific estimator developed for estimating the treatment effect. A key component in constructing the transfer learning estimator is deriving the regression coefficient estimation within each stratum, accounting for the bias between source and target data. To provide a theoretical foundation, we derive the convergence rate for the estimated regression coefficients and subsequently establish the asymptotic normality for the transfer learning estimator. Our results show that when external trial data resemble current trial data, the sample size requirements can be reduced compared to using only current trial data. Finally, we propose a consistent nonparametric variance estimator to facilitate inference that is robust to model misspecifications and applicable to various commonly used randomization procedures. Numerical studies demonstrate the effectiveness and robustness of our proposed estimator across various scenarios. Our study highlights the potential of transfer learning in analyzing randomized clinical trials.
JMLR Optimization & Learning Theory · 发表于 2026
Kelly W. Zhang, Thomas Baldwin-McDonald, Kamil Ciosek, Lucas Maystre, Daniel Russo
摘要
Increasingly, recommender systems are tasked with improving users' long-term satisfaction. In this context, we study a content exploration task, which we formalize as a bandit problem with delayed rewards. There is an apparent trade-off in choosing the learning signal: waiting for the full reward to become available might take several weeks, slowing the rate of learning, whereas using short-term proxy rewards reflects the actual long-term goal only imperfectly. First, we develop a predictive model of delayed rewards that incorporates all information obtained to date. Rewards as well as shorter-term surrogate outcomes are combined through a Bayesian filter to obtain a probabilistic belief. Second, we devise a bandit algorithm that quickly learns to identify content aligned with long-term success using this new predictive model. We prove a regret bound for our algorithm that depends on the Value of Progressive Feedback, an information-theoretic metric that captures the quality of short-term leading indicators that are observed prior to the long-term reward. We apply our approach to a podcast recommendation problem, where we seek to recommend shows that users engage with repeatedly over two months. We empirically validate that our approach significantly outperforms methods that optimize for short-term proxies or rely solely on delayed rewards, as demonstrated by an A/B test in a recommendation system that serves hundreds of millions of users.
JMLR Latent Variables & Mixtures Causal Inference · 发表于 2026
Yuhang Liu, Zhen Zhang, Dong Gong, Mingming Gong, Biwei Huang, Anton van den Hengel, Kun Zhang, Javen Qinfeng Shi
摘要
The task of causal representation learning aims to uncover latent higher-level causal variables that affect lower-level observations. Identifying the true latent causal variables from observed data, while allowing instantaneous causal relations among latent variables, remains a challenge, however. To this end, we start with the analysis of three intrinsic indeterminacies in identifying latent variables from observations: transitivity, permutation indeterminacy, and scaling indeterminacy. We find that transitivity acts as a key role in impeding the identifiability of latent causal variables. To address the unidentifiable issue due to transitivity, we introduce a novel identifiability condition where the underlying latent causal model satisfies a linear-Gaussian model, in which the causal coefficients and the distribution of Gaussian noise are modulated by an additional observed variable. Under certain assumptions, including the existence of a reference condition under which latent causal influences vanish, we can show that the latent causal variables can be identified up to trivial permutation and scaling, and that partial identifiability results can still be obtained when this reference condition is violated for a subset of latent variables. Furthermore, based on these theoretical results, we propose a novel method, termed Structural caUsAl Variational autoEncoder (SuaVE), which directly learns causal representations and causal relationships among them, together with the mapping from the latent causal variables to the observed ones. Experimental results on synthetic and real data demonstrate the identifiability and consistency results and the efficacy of SuaVE in learning causal representations.
JMLR Causal Inference Networks & Graphs · 发表于 2026,收录于 2026-11-06
Vincent Guan, Joseph Janssen, Hossein Rahmani, Andrew Warren, Stephen Y. Zhang, Elina Robeva, Geoffrey Schiebinger
摘要
Stochastic differential equations (SDEs) are a fundamental tool for modelling dynamic processes, including gene regulatory networks (GRNs), contaminant transport, financial markets, and image generation. However, learning the underlying SDE from data is a challenging task, especially if individual trajectories are not observable. Motivated by burgeoning research in single-cell datasets, we present the first comprehensive approach for jointly identifying the drift and diffusion of an SDE from its temporal marginals. Assuming linear drift and additive diffusion, we show that non-identifiability can only arise if the initial distribution possesses generalized rotational symmetries. We further prove that even if this condition holds, the drift and diffusion can almost always be recovered from the marginals. Additionally, we show that the causal graph of any SDE with additive diffusion can be recovered from the identified SDE parameters. To complement this theory, we adapt entropy-regularized optimal transport to handle anisotropic diffusion, and introduce APPEX (Alternating Projection Parameter Estimation from X0), an iterative algorithm designed to estimate the drift, diffusion, and causal graph of an additive noise SDE, solely from temporal marginals, and demonstrate its effectiveness on simulated data from linear additive noise SDEs.
Econometrica Causal Inference · 发表于 2026
Manu Navjeevan, Rodrigo Pinto, Andres Santos
摘要
We study instrumental variables models characterized by: (i) Unobserved heterogeneity consisting of potential outcomes and response types that describe how the instrument determines treatment choice; (ii) Conditional independence of the instrument and the unobserved heterogeneity; and (iii) Convex restrictions on the distribution of unobserved heterogeneity. We show certain causal parameters are identified in these models if and only if a version of the kappa of Abadie (2003) exists. Our identification results are constructive in yielding estimating moment conditions. Focusing on a leading special case, we develop asymptotically normal estimators based on a doubly robust version of these moment conditions.
JMLR Causal Inference Missing Data & Measurement Error · 发表于 2026
Shuozhi Zuo, Peng Ding, Fan Yang
摘要
Under the instrumental variable model, we can identify the local average treatment effect, also known as the complier average causal effect (CACE). In practice, however, the treatment and outcome are often missing. When they are missing not at random (MNAR), the underlying data distribution cannot be recovered, so the CACE is generally not identifiable without further assumptions. We study the conditions under which the CACE remains identifiable when data are MNAR. Searching exhaustively over missingness mechanisms, we characterize those that identify the CACE without auxiliary information in two settings: (1) missing data in the treatment alone or the outcome alone, and (2) missing data in both the treatment and the outcome under prospective data collection. We unify existing results and establish new ones, giving a complete picture of identifiability in each setting. The results have two practical implications. First, before analyzing data under the instrumental variable model, one should check whether the CACE is identifiable under the assumed missingness mechanism. Second, because the true mechanism is usually unknown and untestable, it is more robust to estimate the CACE under several plausible mechanisms and compare the results.
Econometrica Econometrics & Economics · 发表于 2026
Nicholas Ryan
摘要
This paper studies how the risk of hold‐up affects procurement. I use data on the universe of solar power auctions in India. The Indian context allows clean estimates of counterparty risk, because solar plants set up in the same states, by the same firms, are procured in auctions intermediated by either risky states themselves or the trusted central government. I find that the counterparty risk of an average state increases solar prices by 10%. This risk premium sharply reduces investment, because demand for green energy is elastic. Contract intermediation by the central government eliminates the counterparty risk premium.
JMLR High-Dimensional · 发表于 2026
Zelin He, Ying Sun, Jingyuan Liu, Runze Li
摘要
Parameter transfer aims to improve parameter estimation accuracy by leveraging knowledge from related sources. This paper studies the parameter transfer problem from heterogeneous sources for high-dimensional M-estimators. Specifically, we propose a novel one-step estimator with a fused-regularizer and a target-data-oriented constraint, which can robustly capture parameter knowledge from source data in the presence of different types of data distribution shifts. Nonasymptotic bound is provided for the estimation error of target parameter, showing the proposed estimator could achieve effective parameter transfer under distribution shifts, and is guaranteed to perform no worse than any estimators learned only from the target data. We further show that the proposed estimator can achieve the minimax-optimal rate under much weaker conditions than existing methods. In addition, we extend the method to a distributed setting, requiring just one round of communication with source parameter estimators, while retaining the estimation accuracy of the centralized version. Extensive simulations and real data analysis further verify the effectiveness of the method.
JMLR Networks & Graphs High-Dimensional Machine Learning · 发表于 2026
Simon Martin, Giulio Biroli, Francis Bach
摘要
We study the high-dimensional training dynamics of a shallow neural network with quadratic activation in a teacher--student setup. We focus on the extensive-width regime, where the teacher and student network widths scale proportionally with the input dimension, and the sample size grows quadratically. This scaling aims to describe overparameterized neural networks in which feature learning still plays a central role. In the high-dimensional limit, we derive a dynamical characterization of the gradient flow, in the spirit of dynamical mean-field theory (DMFT). Under $\ell_2$-regularization, we analyze these equations at long times and characterize the performance and spectral properties of the resulting estimator. This result provides a quantitative understanding of the effect of overparameterization on learning and generalization, and reveals a double descent phenomenon in the presence of label noise, where generalization improves beyond interpolation. In the small regularization limit, we obtain an exact expression for the perfect recovery threshold as a function of the network widths, providing a precise characterization of how overparameterization influences recovery.
JMLR Causal Inference Networks & Graphs · 发表于 2026
Eli N. Weinstein, David M. Blei
摘要
Causal questions often arise in settings where data are hierarchical: subunits are nested within units. Consider students in schools, cells in patients, or cities in states. In these settings, unit-level variables (e.g., a school's budget) may affect subunit-level outcomes (e.g., student test scores), and subunit-level characteristics may aggregate to influence unit-level outcomes. In this paper, we show how to analyze hierarchical data for causal inference. We introduce hierarchical causal models, which extend structural causal models and graphical models by incorporating inner plates to represent nested data structures. We develop a graphical identification technique for these models that generalizes do-calculus. We show that hierarchical data can enable causal identification even when it would be impossible with non-hierarchical data--for example, when only unit-level summaries are available. We develop estimation strategies, including using hierarchical Bayesian models. We illustrate our results in simulation and through a reanalysis of the classic "eight schools" study.
JMLR Nonparametric & Semiparametric · 发表于 2026
Pratik Rathore, Zachary Frangella, Jiaming Yang, Michał Dereziński, Madeleine Udell
摘要
Kernel ridge regression (KRR) is a fundamental computational tool, appearing in problems that range from computational chemistry to health analytics, with a particular interest due to its starring role in Gaussian process regression. However, full KRR solvers are challenging to scale to large datasets: both direct (e.g., Cholesky decomposition) and iterative methods (e.g., PCG) incur prohibitive computational and storage costs. The standard approach to scale KRR to large datasets chooses a set of inducing points and solves an approximate version of the problem, inducing points KRR. However, the resulting solution tends to have worse predictive performance than the full KRR solution. In this work, we introduce a new solver, ASkotch, for full KRR that provides better solutions faster than state-of-the-art solvers for full and inducing points KRR. ASkotch is a scalable, accelerated, iterative method for full KRR that provably obtains linear convergence. Under appropriate conditions, we show that ASkotch obtains condition-number-free linear convergence. This convergence analysis rests on the theory of ridge leverage scores and determinantal point processes. ASkotch outperforms state-of-the-art KRR solvers on a testbed of 23 large-scale KRR regression and classification tasks derived from a wide range of application domains, demonstrating the superiority of full KRR over inducing points KRR. Our work opens up the possibility of as-yet-unimagined applications of full KRR across a number of disciplines.
JMLR Optimization & Learning Theory High-Dimensional · 发表于 2026
Tong Wu
摘要
Tensors, which give a faithful and effective representation to deliver the intrinsic structure of multi-dimensional data, play a crucial role in an increasing number of signal processing and machine learning problems. However, tensor data are often accompanied by arbitrary signal corruptions, including missing entries and sparse noise. A fundamental challenge is to reliably extract the meaningful information from corrupted tensor data in a statistically and computationally efficient manner. This paper develops a scaled gradient descent (ScaledGD) algorithm to directly estimate the tensor factors with tailored spectral initializations under the tensor-tensor product (t-product) and tensor singular value decomposition (t-SVD) framework. With tailored variants for tensor robust principal component analysis, (robust) tensor completion and tensor regression, we theoretically show that ScaledGD achieves linear convergence at a constant rate that is independent of the condition number of the ground truth low-rank tensor, while maintaining the low per-iteration cost of gradient descent. To the best of our knowledge, ScaledGD is the first algorithm that provably has such properties for low-rank tensor estimation with the t-SVD. Finally, numerical examples are provided to demonstrate the efficacy of ScaledGD in accelerating the convergence rate of ill-conditioned low-rank tensor estimation in a number of applications.
JMLR Networks & Graphs Latent Variables & Mixtures Nonparametric & Semiparametric · 发表于 2026
Wenlong Lyu, Yuheng Jia, Hui Liu, Junhui Hou
摘要
The well-known graph-based clustering methods, including spectral clustering, symmetric non-negative matrix factorization, and doubly stochastic normalization, can be viewed as relaxations of the kernel k-means approach. However, we posit that these methods excessively relax their inherent low-rank, nonnegative, doubly stochastic, and orthonormal constraints to ensure numerical feasibility, potentially limiting their clustering efficacy. In this paper, guided by our systematic theoretical analyses, we propose Low-Rank Doubly stochastic clustering (LoRD), a model that only relaxes the orthonormal constraint to derive a probabilistic clustering results. Furthermore, by theoretically establishing the equivalence between orthogonality and Block diagonality under the doubly stochastic constraint, we propose B-LoRD. By integrating block diagonal regularization into LoRD, expressed as the maximization of the Frobenius norm, we enhance clustering performance. To ensure numerical solvability, we transform the non-convex doubly stochastic constraint into a linear convex constraint through the introduction of a class probability parameter. The theoretical demonstration of the gradient Lipschitz continuity of our LoRD and B-LoRD enables the proposal of a projected gradient algorithm whose exact iteration admits a sublinear convergence-rate bound and ensures first-order stationarity of every accumulation point for the exact projected gradient iteration. Extensive experiments underscore the effectiveness of our approaches. The code is publicly available at https://github.com/lwl-learning/LoRD.
JMLR High-Dimensional · 发表于 2026
Felix Benning, Leif Döring
摘要
We prove that all 'gradient span algorithms' have asymptotically deterministic behavior on scaled Gaussian random functions as the dimension tends to infinity. This is a functional generalization of similar results for random quadratic functions and spin glasses. They explain the counterintuitive phenomenon that different training runs of many large machine learning models result in approximately equal cost curves despite random initialization on a complicated non-convex landscape. This 'predictable progress' phenomenon is exploited by the AutoML community: Since the optimization progress of a single run is already representative, multiple retries with the same hyperparameters are not necessary.
JMLR Bayesian Methods · 发表于 2026,收录于 2026-09-10
Javier Burroni, Daniel Sheldon
摘要
Mixture distributions are expressive variational families for black-box VI, but their discrete component choices complicate gradient estimation. We systematize reparameterization-based estimators for mixtures in a common notation, giving self-contained derivations and extending several to new settings. In particular, we provide an elementary derivation of a single-sample post-stratified estimator---previously derived via transport equations---and prove a variance reduction relative to simple random sampling. We also broaden the applicability of implicit reparameterization and reduce its computational complexity. Across different benchmarks, we find that stratified estimators are consistently robust when feasible; among single-sample methods, the post-stratified estimator frequently perform best, while implicit reparameterization is the most computationally demanding. Our analysis clarifies when each method should be used and provides efficient algorithms that make mixture-based variational inference practical.
JMLR Optimization & Learning Theory · 发表于 2026
Álvaro Gajardo, Hans-Georg Müller
摘要
We study manifold learning with multidimensional scaling for samples of metric space valued data. By adopting a global version of ISOMAP we obtain low-dimensional Euclidean representations. A key innovation is that we demonstrate that global Fréchet regression can be utilized for mapping the elements of a convex set in the Euclidean representation space back to the metric space where the objects reside. We refer to this approach as Fréchet manifold learning and showcase it with one-dimensional distributions as random objects, equipped with the Wasserstein metric, which is an important special case of our general approach. The resulting low-dimensional representations mimic the parametric representation in a parametric family of distributions but are entirely learned from the data without postulating any parametric model. These Wasserstein representations of distributional data can be viewed as an empirical parametrization of a sample of distributions. The utility of these representations rests on the map from the low-dimensional Euclidean representation space to the space of distributions, which is obtained with global Fréchet regression. We illustrate the proposed approach with distributional data for baby names, bike rentals and age pyramids and further demonstrate how it can be applied for a novel distributional regression method that features one-dimensional distributions as predictors.
Econometrica Genetics & Omics Econometrics & Economics · 发表于 2026,收录于 2026-10-20
Eduardo Azevedo, Jonathan P. Beauchamp, Richard Karlsson Linnér
摘要
Recent advances have substantially improved our ability to predict disease risk with genetic data. In response, many countries have banned insurers from using genetic data, despite concerns about adverse selection. This paper measures adverse selection in a market where insurers can underwrite only on nongenetic information while consumers also have access to genetic information. We do so by combining methods from quantitative genetics, selection measures from economic theory, and genetic and electronic health record data on nearly 500,000 individuals in the UK Biobank. We focus on the critical illness insurance market and consider scenarios in which consumers have access to current or expected future genetic prediction technology. We find noticeable levels of selection with current prediction technology, and potentially crippling selection with expected future technology.
JMLR Bayesian Methods · 发表于 2026
Yuexi Wang, Veronika Rockova
摘要
In the absence of explicit or tractable likelihoods, Bayesians often resort to approximate Bayesian computation (ABC) for inference. Our work bridges ABC with deep neural implicit samplers based on generative adversarial networks (GANs) and adversarial variational Bayes. Both ABC and GANs compare aspects of observed and fake data to simulate from posteriors and likelihoods, respectively. We develop a Bayesian GAN (B-GAN) sampler that directly targets the posterior by solving an adversarial optimization problem. B-GAN is driven by a deterministic mapping learned on the ABC reference by conditional GANs. Once the mapping has been trained, iid posterior samples are obtained by filtering noise at a negligible additional cost. We propose two post-processing local refinements using (1) data-driven proposals with importance reweighting, and (2) variational Bayes. We support our findings with frequentist-Bayesian results, showing that the typical total variation distance between the true and approximate posteriors converges to zero for certain neural network generators and discriminators. Our findings on simulated data show highly competitive performance relative to some of the most recent likelihood-free posterior simulators.
JMLR Other · 发表于 2026
Diego Marcondes, Ulisses Braga-Neto
摘要
We propose generalized resubstitution error estimators for regression. Each error estimator in this class corresponds to a choice of an empirical probability measure and a loss function. The standard empirical probability measure and the quadratic loss lead to the standard sum of squares error estimator. Other choices of empirical probability measure lead to more general estimators with superior bias and variance properties. We prove that these error estimators are consistent under broad assumptions. In addition, procedures for choosing the empirical measure based on the method of moments and maximum pseudo-likelihood are proposed and investigated. Detailed experimental results using polynomial regression demonstrate empirically the superior finite-sample bias and variance properties of the proposed estimators. The R code for the experiments is provided.
JMLR Optimization & Learning Theory · 发表于 2026,收录于 2026-09-17
Aaron Anderson, Michael Benedikt
摘要
We consider the relationship between learnability of a "base class" of functions on a set X, and learnability of a class of statistical functions derived from the base class. For example, we refine results showing that learnability of a family h_p: p ∈ Y of functions implies learnability of the family of functions h_μ = λp: Y. E_μ(h_p), where E_μ is the expectation with respect to μ, and μ ranges over probability distributions on X. We will look at both Probably Approximately Correct (PAC) learning, where example inputs and outputs are chosen at random, and online learning, where the examples are chosen adversarily. For agnostic learning, we establish improved bounds on the sample complexity of learning for statistical classes, stated in terms of combinatorial dimensions of the base class. We connect these problems to techniques introduced in model theory for "randomizing a structure". We also provide counterexamples for realizable learning, in both the PAC and online settings." (JMLR's own per-paper page was blocked; retrieved from the paper's arXiv mirror, arXiv:2504.00847.) *Publication/online date: date unknown (JMLR page inaccessible; listed in Volume 27, 2026)*
JMLR Other · 发表于 2026,收录于 2026-09-10
Łukasz Dębowski
摘要
We inspect the deductive connection between the neural scaling law and Zipf's law--two statements discussed in machine learning and quantitative linguistics. The neural scaling law describes how the cross entropy rate of a foundation model--such as a large language model--changes with respect to the amount of training tokens, parameters, and compute. By contrast, Zipf's law posits that the distribution of tokens exhibits a power law tail. Whereas similar claims have been made in more specific settings, we show that the neural scaling law is a consequence of Zipf's law under certain broad assumptions that we reveal systematically. The derivation steps are as follows: We derive Heaps' law on the vocabulary growth from Zipf's law, Hilberg's hypothesis on the entropy scaling from Heaps' law, and the neural scaling from Hilberg's hypothesis. We illustrate these inference steps by a toy example of the Santa Fe process that satisfies all four statistical laws.
Econometrica Econometrics & Economics · 发表于 2026
Allan Hsiao, Jacob Moscona, Karthik A. Sastry
摘要
This paper studies how governments intervene in agricultural markets to reshape the economic consequences of climate extremes. We construct a global dataset of agricultural policies and extreme heat exposure by country and crop since 1980. Extreme heat shocks to domestic production lead to policies that assist consumers by lowering domestic food prices. This effect is persistent, primarily implemented via border policies, and stronger during election years. Shocks to foreign production induce the opposite response: policies that assist producers by raising prices. These findings can be rationalized by a model in which governments use agricultural policy to redistribute among domestic interest groups. Our estimates imply that policy responses shield domestic consumers, while exacerbating losses for domestic producers and foreign consumers. Policy responses have regressive consequences globally, disproportionately harming poor and heat‐exposed countries.
JMLR Networks & Graphs Machine Learning · 发表于 2026,收录于 2026-11-06
Alexander Kagan, Elizaveta Levina, Ji Zhu
摘要
Modeling information spread through a network is one of the key problems of network analysis. Most approaches assume the spread is governed by some probabilistic diffusion model, often parameterized by the strength of connections between network members (edge weights), highlighting the need for methods that can accurately estimate them. Multiple prior works suggest estimators for particular diffusion models, but lack a rigorous statistical analysis establishing their asymptotic properties and allowing for uncertainty quantification. In this paper, we develop a likelihood-based approach to estimate edge weights from the observed information diffusion paths under the proposed General Linear Threshold (GLT) model, a broad class of discrete-time information diffusion models that includes both the well-known linear threshold (LT) and independent cascade (IC) models. We derive a finite-sample error bound for the estimator and demonstrate that it is asymptotically normal under mild conditions. We conclude by studying the GLT model in the context of the Influence Maximization (IM) problem. Extensive experiments on synthetic and real-world networks demonstrate that the proposed estimation and inference frameworks for the GLT model parameters can significantly improve estimation of spread from a given subset of nodes, prediction of node activation, and the quality of the IM problem solutions.
JMLR Causal Inference · 发表于 2026
Jiayi Wang, Raymond K. W. Wong, Xiaoke Zhang, Kwun Chuen Gary Chan
摘要
We study treatment effect estimation with functional treatments where the average potential outcome functional is a function of functions, in contrast to continuous treatment effect estimation where the target is a function of real numbers. By considering a flexible scalar-on-function marginal structural model, a weight-modified kernel ridge regression (WMKRR) is adopted for estimation. The weights are constructed by directly minimizing the uniform balancing error resulting from a decomposition of the WMKRR estimator, instead of being estimated under a particular treatment selection model. Despite the complex structure of the uniform balancing error derived under WMKRR, finite-dimensional convex algorithms can be applied to efficiently solve for the proposed weights thanks to a representer theorem. The optimal convergence rate is shown to be attainable by the proposed WMKRR estimator without any smoothness assumption on the true weight function. Corresponding empirical performance is demonstrated by a simulation study and a real data application.
JMLR Networks & Graphs Machine Learning · 发表于 2026
Nikolaos Tsilivis, Eitan Gronich, Julia Kempe, Gal Vardi
摘要
We study the implicit bias of the general family of steepest descent algorithms with infinitesimal learning rate in deep homogeneous neural networks. We show that: (a) an algorithm-dependent geometric margin starts increasing once the networks reach perfect training accuracy, and (b) any limit point of the training trajectory corresponds to a KKT point of the corresponding margin-maximization problem. We experimentally zoom into the trajectories of neural networks optimized with various steepest descent algorithms, highlighting connections to the implicit bias of popular adaptive methods (Adam and Shampoo).
Econometrica Econometrics & Economics · 发表于 2026
Kate Ho, Robin S. Lee
摘要
We investigate how formularies used by pharmacy benefit managers (PBMs) affect equilibrium manufacturer rebates for branded drugs through tiering and exclusion. We develop a theoretical model of multidimensional contracting in which a PBM negotiates with drug manufacturers over menus of formulary‐contingent rebates and chooses a formulary. We then estimate consumer demand responses to tier placement for statins using claims data from Princeton University, a large employer contracting with a single PBM to offer prescription drug coverage to its employees. Combining the theoretical model with demand estimates and observed list prices, we quantify how allowing for differential tier placement and exclusion affect equilibrium rebates. Our predictions are consistent with available aggregate rebate data, and we find that allowing a PBM to place branded drugs on preferred‐ and non‐preferred tiers can substantially increase negotiated rebate payments.
Econometrica Econometrics & Economics · 发表于 2026
Jonathan Eaton, Samuel Kortum, Francis Kramarz
摘要
Customs data reveal the heterogeneity and granularity of relationships among buyers and sellers, showing how more exports to a destination break down into more firms selling there and more buyers per exporter. We develop a quantitative general equilibrium model of firm‐to‐firm matching that builds on this insight to separate the roles of iceberg costs and matching frictions in gravity. In the cross section, we find matching frictions as important as iceberg costs in impeding trade, and more sensitive to distance. Because domestic and imported intermediates compete directly with labor in performing production tasks, our model also fits the heterogeneity of labor shares across French producers. Applying the framework to the 2004 expansion of the European Union, reduced iceberg costs and reduced matching frictions contributed equally to the increase in French exports to the new members. While workers benefited overall, those competing most directly with imports gained less, even losing in some countries entering the EU.
Econometrica Econometrics & Economics · 发表于 2026
Naoki Aizawa, Corina Mommaerts, Stephanie Rennane
摘要
This paper studies the labor market impacts of firm accommodation decisions after workplace disability and assesses implications for the design of firm subsidies. We leverage a workers' compensation (WC) program in Oregon that provides wage subsidies to firms for accommodating workers with workplace disabilities. Leveraging rich administrative data and a policy change to the wage subsidy, we show that accommodation rates respond to the subsidy rate and that receipt of accommodation leads to a significant increase in employment and earnings a year later. To explore welfare implications, we develop and estimate a frictional labor market model of accommodation as a form of human capital investment. Worker turnover and imperfect experience rating in WC lead to underaccommodation and inefficient labor market outcomes after workplace disability. Counterfactual simulations show that subsidizing accommodation not only improves long‐run labor market outcomes of workers experiencing work‐related disability but also yields welfare gains for most workers.
JMLR Other · 发表于 2026
Yuze Han, Xiang Li, Zhihua Zhang
摘要
In two-time-scale stochastic approximation (SA), two iterates are updated at varying speeds using different step sizes, with each update influencing the other. Previous studies on linear two-time-scale SA have shown that the convergence rates of the mean-square errors for these updates depend solely on their respective step sizes, a phenomenon termed decoupled convergence. However, achieving decoupled convergence in nonlinear SA remains less understood. Our research investigates the potential for finite-time decoupled convergence in nonlinear two-time-scale SA. We demonstrate that, under a nested local linearity assumption, finite-time decoupled convergence rates can be achieved with suitable step size selection. To derive this result, we conduct a convergence analysis of the matrix cross term between the iterates and leverage fourth-order moment convergence rates to control the higher-order error terms induced by local linearity. To further investigate the necessity of local linearity for decoupled convergence, we also construct an example showing that, even when the fast-time-scale update is linear, the nonlinearity of the slow-time-scale update alone can destroy decoupled convergence.
JMLR Machine Learning Networks & Graphs Bayesian Methods · 发表于 2026
Steven Adams, Andrea Patanè, Morteza Lahijanian, Luca Laurenti
摘要
Infinitely wide or deep neural networks (NNs) with independent and identically distributed (i.i.d.) parameters have been shown to be equivalent to Gaussian processes. Because of the favorable properties of Gaussian processes, this equivalence is commonly employed to analyze neural networks and has led to various breakthroughs over the years. However, neural networks and Gaussian processes are equivalent only in the limit; in the finite case there are currently no methods available to approximate a trained neural network with a Gaussian model with bounds on the approximation error. In this work, we present an algorithmic framework to approximate a neural network of finite width and depth, and with not necessarily i.i.d. parameters, with a mixture of Gaussian processes with bounds on the approximation error. In particular, we consider the Wasserstein distance to quantify the closeness between probabilistic models and, by relying on tools from optimal transport and Gaussian processes, we iteratively approximate the output distribution of each layer of the neural network as a mixture of Gaussian processes. Crucially, for any NN and $\epsilon >0$ our approach is able to return a mixture of Gaussian processes that is $\epsilon$-close to the NN at a finite set of input points. Furthermore, we rely on the differentiability of the resulting error bound to show how our approach can be employed to tune the parameters of a NN to mimic the functional behavior of a given Gaussian process, e.g., for prior selection in the context of Bayesian inference. We empirically investigate the effectiveness of our results on both regression and classification problems with various neural network architectures. Our experiments highlight how our results can represent an important step towards understanding neural network predictions and formally quantifying their uncertainty.
JMLR Testing & Uncertainty · 发表于 2026,收录于 2026-10-08
Lin Lu, Yuyang Huo, Haojie Ren, Zhaojun Wang, Changliang Zou
摘要
*(abstract from the JMLR article page, verbatim)* This work studies online multiple testing with feedback, where decisions are made sequentially, and the true state of the hypothesis is revealed after decisions are made, either instantly or with a delay, and under either full or bandit feedback. We propose Generalized alpha-investing with feedback (GAIF) along with its adaptive variants, a feedback-enhanced framework that dynamically adjusts thresholds using revealed outcomes, ensuring finite-sample false discovery rate (FDR)/marginal FDR (mFDR) control. We further extend GAIF to online conformal testing by constructing valid conformal $p$-values and developing feedback-enhanced testing rules with finite-sample mFDR control. We also propose a feedback-driven score selection criterion to adaptively choose the candidate score that is most effective for the testing procedure, together with a theoretical analysis of its optimality. Numerical simulations and real-data applications demonstrate the effectiveness of our methods.
JMLR Networks & Graphs Spatial & Time Series · 发表于 2026
Samuel Cognolato, Alessandro Sperduti, Luciano Serafini
摘要
The Deep Graph Generation's panorama spans two extremes: one-shot and sequential models. The former generates nodes and edges jointly, while the latter samples them autoregressively. Each method performs better in different graph domains depending on size and topology, but neither is applicable to all graph categories. For instance, one-shot methods struggle with generating large graphs, while sequential methods underperform on smaller graphs. A possible way to overcome these limitations is to flexibly combine the two methods in a unique system. In this work, we propose the FLAGG (Flexible Autoregressive Graph Generation) framework, which sequentially generates portions of graphs with one-shot models. FLAGG can apply any one-shot model to make it autoregressive, allowing flexibility in choosing the sequential policy. This policy is specified through a stochastic node removal process, which an Insertion Model learns to reverse. We evaluate FLAGG with the DiGress one-shot model on several data sets of different graph sizes and domains. We show that the approach outperforms both one-shot and autoregressive baselines in terms of sampling quality.
JMLR Bayesian Methods High-Dimensional · 发表于 2026
Bohan Wu, David M. Blei
摘要
Variational inference (VI) has emerged as a popular method for approximate inference for high-dimensional Bayesian models. In this paper, we propose a novel VI method that extends the naive mean field via entropic regularization, referred to as $\Xi$-variational inference ($\Xi$-VI). $\Xi$-VI has a close connection to the entropic optimal transport problem and benefits from the computationally efficient Sinkhorn algorithm. We show that $\Xi$-variational posteriors effectively recover the true posterior dependency, where the likelihood function is downweighted by a regularization parameter. We analyze the role of dimensionality of the parameter space on the accuracy of $\Xi$-variational approximation and the computational complexity of computing the approximate distribution, providing a rough characterization of the statistical-computational trade-off in $\Xi$-VI, where higher statistical accuracy requires greater computational effort. We also investigate the frequentist properties of $\Xi$-VI and establish results on consistency, asymptotic normality, high-dimensional asymptotics, and algorithmic stability. We provide sufficient criteria for our algorithm to achieve polynomial-time convergence. Finally, we show the inferential benefits of using $\Xi$-VI over mean-field VI and other competing methods, such as normalizing flow, on simulated and real datasets.
JMLR Survival Analysis · 发表于 2026
Yudong Wang, Zhi-Sheng Ye, Cheng Yong Tang
摘要
Predicting future time-to-event outcomes is a foundational task in statistical learning. While various methods exist for generating point predictions, quantifying the associated uncertainties poses a more substantial challenge. In this study, we introduce an innovative approach specifically designed to address this challenge, accommodating dynamic predictors that may manifest as stochastic processes. Our investigation harnesses the forward intensity function in a novel way, providing a fresh perspective on this intricate problem. The framework we propose demonstrates remarkable computational efficiency, enabling efficient analyses of large-scale investigations. We validate its soundness with theoretical guarantees, and our in-depth analysis establishes the weak convergence of function-valued parameter estimations. We illustrate the effectiveness of our framework with two comprehensive real examples and extensive simulation studies.
JMLR Machine Learning · 发表于 2026,收录于 2026-11-06
Lei Cheng, Wen Shen, Quanshi Zhang, Yuchen Huang, Yiming Zhang, Yuxiao Yang, Mingjie Li
摘要
In this paper, we examine whether a set of sparse symbolic interactions can faithfully explain the complex inference patterns of large language models (LLMs), thereby using interactions to diagnose the detailed representations of LLMs. Since previous studies lack a comprehensive proof for interaction-based explanations, we first experimentally verify the sparsity of interactions in LLMs, i.e., we observe that 50--100 interactions are usually powerful enough to accurately mimic all subtle changes in the LLM's output scores over an exponential number of arbitrary masked inputs. More crucially, interactions provide new perspectives for analyzing an LLM. (1) We propose to disentangle the interactions encoded by the LLM in a coarse-to-fine manner and use interactions to analyze the increasing representational complexity of features during forward propagation. The clear disentanglement of interactions enables us to identify incorrect inference patterns hidden in an LLM. (2) We empirically demonstrate that LLMs with different architectures encode similar interactions for text generation, and simple interactions are more likely to be consistently extracted by different LLMs than complex interactions. (3) We also empirically demonstrate that an LLM encodes similar interactions over semantically similar input sentences. Again, simple interactions are more likely to be stably extracted than complex interactions. (4) We empirically find that low-order interactions can serve as indicators for guiding model compression.
JMLR Machine Learning · 发表于 2026
Tianxing Mei, Yingying Fan, Jinchi Lv
摘要
We offer theoretical and empirical insights into the impact of exogenous randomness on the effectiveness of random forests with tree-building rules independent of training data. We formally introduce the concept of exogenous randomness which can come from feature subsampling or tie-breaking in tree-building processes. We develop non-asymptotic expansions for the mean squared error (MSE) for both individual trees and forests and establish sufficient and necessary conditions for their consistency. In the special example of the linear regression model with independent features, our MSE expansions are more explicit, providing more understanding of the random forests' mechanisms. It also allows us to derive an upper bound on the MSE with explicit consistency rates for trees and forests. Guided by our theoretical findings, we conduct simulations to further explore how exogenous randomness enhances random forests performance. Our findings unveil that feature subsampling reduces both the bias and variance of random forests compared to individual trees, serving as an adaptive mechanism to balance bias and variance. Furthermore, our results reveal an intriguing phenomenon: the presence of noise features can act as a “blessing" in enhancing the performance of random forests thanks to feature subsampling.
Econometrica Econometrics & Economics · 发表于 2026
Battal Dogan, Lars Ehlers
摘要
In the context of one‐to‐one matching markets, we study myopic‐farsighted stable sets , which are internally and externally stable when myopic agents consider immediate payoffs from their deviations, while farsighted agents anticipate counter‐deviations and consider final payoffs. We constructively prove the existence of a (rational expectations) myopic‐farsighted stable set, in which farsighted agents receive a single payoff while myopic agents may receive multiple payoffs. Our existence result extends to settings with enforcing coalitions of arbitrary size, yielding coalitional myopic‐farsighted stable sets , and to settings where not all members of an enforcing coalition must strictly gain, yielding myopic‐farsighted weakly stable sets . When all farsighted agents have unit demand, our results also extend to many‐to‐one matching markets. As a key corollary, we provide a foundation for the efficiency‐adjusted deferred acceptance algorithm by showing that its outcome constitutes a singleton myopic‐farsighted stable set when one side is farsighted and the other is myopic.