Clay weekly context brief for the Statistics category (ISO week 2026-W37).

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. Semiparametric Estimation of Delayed-Outcome Treatment Effects Using Short-Term Surrogates under Administrative Censoring
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2603.10405
   The multi-site registry studies, such as Stepped-wedge cluster-randomized trials (SW-CRT), staggered-enrollment RCTs, etc., share a structural feature: the primary long-term outcome is administratively censored for a non-negligible fraction of units, with censoring driven by calendar design rather than by the outcome itself.
2. Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.02899
   Benchmark contamination, the leakage of test items into training data, is widely described as a threat to the reliability of large language model (LLM) leaderboards.
3. Co-SIVI: A Correlated Semi-Implicit Variational Approach for Spatial Models
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2510.19722
   We propose correlated semi-implicit variational inference (Co-SIVI), a scalable approach for full posterior approximation in large spatial models with exponential-family likelihoods.
4. Change-point analysis: a new perspective for unstable financial markets
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.03614
   We introduce two new classes of nonparametric change-point tests for sequences of univariate non-negative random variables.
5. A Closed-Form Formula for Consistent Lipschitz Regression on Metric Spaces with Sparse Neural Network Realizations
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.03129
   Several classical machine-learning methods, such as KRRs and SVRs, are both computationally and analytically tractable since their estimators either admit closed-form expressions or are obtained by minimizing convex training objectives; neither feature is generally available for deep neural networks.
6. Induction and the rule of succession through a possibilistic inferential model lens
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2608.11935
   Induction is the process by which empirical evidence is transformed to knowledge.
7. Distributional Treatment Effect Transportability across Heterogeneous Sites
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2511.09759
   We study distributional transportability of treatment effects in a ``cross-site, one-armed target" design, where both treated and control units are observed in a source site, but only control units are observed in a target site.
8. SafeRestore: Detector-Relative Risk Certificates for Selective Industrial Image Restoration
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.03475
   Industrial inspection pipelines often restore a measured image before a detector acts on it, yet restoration can suppress detector-supported defect structure or create clean-region activations.
9. Markov Chain Monte Carlo with Diffusion Paths
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2607.11631
   Sampling from multimodal distributions is a longstanding challenge for classical local Markov chain Monte Carlo (MCMC) methods.
10. Parametric estimation of Hawkes processes based on ordinary least squares
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.03696
   We develop a parametric estimation framework for self-exciting Hawkes processes whose intensity functions admit a parametric form.
11. ALRA: Adaptive Local Relational Alignment for Logit-Based Pre-training Distillation of Autoregressive Language Models
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.03355
   Logit-based knowledge distillation for autoregressive language models usually aligns teacher and student next-token distributions over the entire vocabulary.
12. Fully specified Bayes factors for hypothesis testing and sensitivity analysis in process tracing
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2606.16683
   In process tracing, researchers ask how strongly their evidence favors their explanation, the working theory, over a rival.
13. Geometric Optics Approximation Sampling: A Reflector-Induced Transport Map Framework
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2403.01655
   In this paper, we propose Geometric Optics Approximation Sampling (GOAS), a reflector-induced transport-map framework for sampling from target measures.
14. Random mixtures in Bayes Hilbert spaces
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.03523
   We present a framework for the analysis and unmixing of random density mixtures in the Bayes Hilbert space.
15. Modeling uncertainty in the covariance matrix for probabilistic forecast reconciliation
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2506.19554
   In minimum trace (MinT) forecast reconciliation, the covariance matrix of the base forecast errors plays a crucial role.
16. Comment on: "The Two Cultures of Prevalence Mapping: Small Area Estimation and Model-Based Geostatistics"
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.03805
   Small Area Estimation (SAE) and Model-Based Geostatistics (MBG) provide complementary approaches to prevalence mapping, with their relative advantages depending on the inferential goals and characteristics of the available data.
17. Towards Scaling Reinforcement Learning to Massive Populations: Learning Mean-Field Representations
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.02928
   Modern multi-agent systems are increasingly deployed at scale over large populations of agents in settings such as ad-auctions, traffic routing, and recommendation systems.
18. The Software Behind the Stats: A Student Exploration of Software Trends Across Disciplines
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2504.06507
   This paper presents a student-led activity designed to explore the use of statistical software in academic research across economics, political science, and statistics.
19. Robust dimension-free estimation of simple random tensors: optimal guarantees under heavy tails and adversarial contamination
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2609.00675
   We study robust estimation of simple random tensors of arbitrary order $q\in\mathbb{N}$ under finite-moment assumptions and adversarial contamination.
20. A location-invariant estimator of extremal quantile treatment effects for heavy-tailed distributions
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.04018
   Quantile treatment effects (QTEs) measure the effect of a treatment on the distribution of an outcome, and their estimation at extreme quantile levels is of central interest in applications where the target quantiles lie far beyond the range of the data.
21. Metaorder modelling and identification from public data
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2602.19590
   Market-order flow in financial markets exhibits long-range correlations.
22. Causal Inference for Heterogeneous Extreme Quantiles with Heavy-Tailed Outcomes
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.03933
   We propose a framework for estimating conditional extreme quantile treatment effects (CEQTEs) in observational studies with heavy-tailed outcomes.
23. The Geometry of Ignorance: LLMs Know When to Temper Bayesian Priors
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.02959
   What does a language model predict when it has few clues?
24. Statistical Leadership of What? Statistics After AI
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2608.29629
   Statisticians have spent over a century arguing that we are more than calculators, usually by pointing to what else we know.
25. Gaussian comparison above the median
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2607.06874
   We prove a Gaussian comparison inequality for closed convex sets with reference probability at least 1/2.
26. Assessing the impact of tourist attractions through the integration of causal inference and demand-side economic analysis: A case study of the Sensoria
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2605.21464
   Assessing the economic impact of tourist attractions typically adopts a demand-side economic approach, which is frequently based on visitor surveys.
27. TrunX: A massively parallel, differentiable implementation of the 3-PG forest growth model in JAX
   Source: stat.CO (Computation)
   Link: https://arxiv.org/abs/2609.02557
   Process-based forest models are widely used to simulate forest growth and responses to environmental change, but their calibration and application often require many computationally expensive model evaluations.
28. Model-assisted estimation with a training subsample: a two-phase sampling approach with design-based variance estimation
   Source: stat.ME (Methodology)
   Link: https://arxiv.org/abs/2609.04082
   When a flexible prediction model is fitted on a training subsample drawn from a probability sample, the model-assisted estimator actually reported arises from one realized partition, yet existing theory quantifies uncertainty only for partition-averaged, cross-fitted, or symmetrized versions of it.
29. Tail-Likelihood Reinforcement Learning
   Source: stat.ML (Machine Learning)
   Link: https://arxiv.org/abs/2609.02987
   Reinforcement learning typically optimizes average reward.
30. E-values as statistical evidence: A comparison to Bayes factors, likelihoods, and p-values
   Source: stat.OT (Other Statistics)
   Link: https://arxiv.org/abs/2603.24421
   A recurring debate in the philosophy of statistics concerns what, exactly, should count as a measure of evidence for or against a given hypothesis.
31. A robust and scalable estimation for high-dimensional volatility models
   Source: stat.TH (Statistics Theory)
   Link: https://arxiv.org/abs/2510.17578
   This paper introduces a robust and computationally efficient estimation framework for high-dimensional volatility models in the BEKK-ARCH class.
32. Bigraphical Mat\'ern-Whittle (BMW) Processes for Fast Inference of Big Multivariate Spatial Data on General Domains
   Source: stat.AP (Applications)
   Link: https://arxiv.org/abs/2609.01950
   Large spatial data sets now record many correlated variables at many thousands of locations, often on domains where Euclidean distance misrepresents proximity.

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 418 available items for this weekly brief.