Clay weekly context brief for the Statistics category (ISO week 2026-W40).

Clay tracks publications from the Statistics feed list. Below are recent items from this category, each with its source and a short description of what the publication covers when one is available in the source feed.

Recent publications:
1. RLVR landscapes for iterated multiplications can be benign: Insights from spin-glass theory
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.28625
   Despite the importance of reinforcement learning with verifiable rewards (RLVR), the extent to which it can learn new reasoning capabilities remains debated.
2. Modeling cyclostationarity in time series using ASCA
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2603.05065
   Modern data analysis across diverse disciplines increasingly relies on time series.
3. glmSTARMA -- An R-Package for fitting autoregressive spatio-temporal models following generalized linear models
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2607.08276
   The R package glmSTARMA implements autoregressive models for spatio-temporal data at fixed locations, with time-invariant spatial dependency structure.
4. Auditing Bayesian Graph Alignment: Diagnostic Comparisons and Reference Failure
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.23232
   Bayesian graph alignment estimates correspondence probabilities, but convergence of an alignment-score trace need not imply accurate correspondence marginals.
5. Confidence Horizons
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2608.03889
   Anytime-valid inference enables analysts to continuously monitor their data and stop experiments early.
6. Engaging students with statistics through choice of real data context on homework
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2603.04541
   Statistics educators recommend teaching with real data with relevant contexts, but defining relevancy is challenging and varies by student.
7. SGA: Uncertainty Quantification for Multi-Step Forecasting in Time Series Foundation Models
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.28582
   The recent emergence of Time Series Foundation Models (TSFMs) has significantly advanced multi-step forecasting performance, enabling accurate predictions over extended future horizons.
8. Field Theory of Bayesian Rate Estimation
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2509.06129
   We address the statistical inference of a time-dependent rate of events in the framework of Bayesian field theory.
9. DeepGOF-1: A Pretrained Convolutional Goodness-of-Fit Test for Logistic Regression with a Computable Consistency Certificate
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2609.29575
   Goodness-of-fit tests for logistic regression are least reliable where they are most needed: at small samples their levels drift from the nominal one, and combining them worsens the drift.
10. Do Audio Language Models Hear and Read Distinctive Features Alike?
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.30167
   Audio language models pass speech and text through a single decoder.
11. All you need is log
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2606.27349
   How different are several probability distributions from one another?
12. Asymptotic confidence intervals for the difference and the ratio of the weighted kappa coefficients of two diagnostic tests subject to a paired design
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2407.21387
   The weighted kappa coefficient of a binary diagnostic test is a measure of the beyond-chance agreement between the diagnostic test and the gold standard, and depends on the sensitivity and specificity of the diagnostic test, on the disease prevalence and on the relative importance between the false positives and the false negatives.
13. From Prediction to Explainable Provider Behavior Profiles for Fraud, Waste, and Abuse Review
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.28477
   Claims data can show that provider behavior changed but cannot by itself explain why.
14. Decision Theoretic Subgroup Detection With Bayesian Machine Learning
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2509.05832
   We consider the problem of identifying promising subpopulations in terms of treatment effectiveness or treatment effect heterogeneity, from a Bayesian decision theoretic perspective.
15. Fitting Large Nonlinear Mixed Effects Models Using Variational Expectation Maximization
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2604.26160
   Nonlinear Mixed Effects (NLME) models are widely used in pharmacometrics and related fields to analyze hierarchical and longitudinal data.
16. Learning the Maximum Tolerated Dose for Continuous Toxicity via Monotone Bayesian Trees
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.30190
   Phase I cancer trials seek the maximum tolerated dose (MTD) while protecting patients from excessive toxicity.
17. Pointwise Generalization in Deep Neural Networks
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2605.18598
   We address the fundamental question of why deep neural networks generalize by establishing a pointwise generalization theory for fully connected networks.
18. Inverse Problems Conditioned on Observation Ensembles: Applications and Methods
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2601.22029
   We introduce a new multivariate statistical problem that we refer to as the Ensemble-conditioned Inverse Problem (EIP).
19. Bridging Impulse Control of Piecewise Deterministic Markov Processes and Markov Decision Processes: Frameworks, Extensions, and Open Challenges
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2501.04120
   Control theory plays a pivotal role in understanding and optimizing the behavior of complex dynamical systems across various scientific and engineering disciplines.
20. BLOC: A Global Optimization Framework for Sparse Covariance Estimation with Non-Convex Penalties
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2603.29169
   We introduce BLOC (Black-box Optimization over Correlation matrices), a general framework for sparse covariance estimation with non-convex penalties.
21. Bayesian joint modeling of longitudinal patient-reported outcomes and survival: an application to chronic obstructive pulmonary disease
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.30188
   Questionnaire-based patient-reported outcomes (PROs) are discrete, bounded and overdispersed, yet joint models relating them to survival may ignore these features or estimate both processes sequentially.
22. Stacked SVD or SVD stacked? A Random Matrix Theory perspective on data integration
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2507.22170
   Modern data analysis increasingly requires identifying shared latent structure across multiple high-dimensional datasets.
23. Nuclear Norm-Regularized Bayesian Matrix Completion
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.30078
   Matrix completion, the problem of estimating missing entries in a matrix from noisily observed ones, underlies a diverse array of problems such as recommender systems and counterfactual outcome estimation in panel data.
24. Three Ways Classical Test Theory Misleads for LLM Judges
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.29709
   An LLM judge scores a bank of responses against a rubric, and the reliability comes back at $0.52$.
25. Delicatessen: Automated Estimating Equations in Python
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2203.11300
   Estimating equation theory provides a unified framework for statistical modeling and inference.
26. The risks of dichotomising ordinal outcomes: A spatial analysis of self-rated health in Western Europe
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.29893
   Survey responses are often measured using ordered response categories.
27. Self-Normalizing Denominators in Rational Covariance Estimators
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2608.20223
   Many estimators are ratios of coprime polynomials in a sample covariance matrix, and their accuracy depends on the relative fluctuation of the sample denominator.
28. Path-specific harm decomposition: A partial identification framework
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.29938
   A central goal when designing treatment policies is often to "do no harm", that is, to avoid interventions that improve average outcomes while worsening outcomes for some individuals.
29. A Proxy-likelihood Estimator for Multivariate Extremes Models with Intractable Likelihoods
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.30244
   Many multivariate extremes models have intractable likelihoods requiring practitioners to use alternative fitting methods.
30. BFI: An R Package for Bayesian Federated Inference
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2609.27977
   Bayesian Federated Inference (BFI) estimates statistical models from multicenter data when individual-level observations cannot be combined across centers.
31. Hybrid Models for Short-Term Sea-Level Forecasting
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.29827
   Accurate tide forecasts are essential for coastal management, navigation, flood-risk reduction, and infrastructure protection.
32. Mean Residual Life Ageing Intensity Function
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2409.10456
   Ageing intensity is usually formulated through the failure rate, whereas its mean residual life (\(MRL\))-based counterpart has not been systematically developed.

Sources in this brief: stat.AP (Applications); stat.CO (Computation); stat.ME (Methodology); stat.ML (Machine Learning); stat.OT (Other Statistics); stat.TH (Statistics Theory).
Selected 32 of 408 available items for this weekly brief.