Papers

Modern datasets arising from social media, genomics, and biomedical informatics are often heterogeneous and (ultra) high-dimensional, creating substantial challenges for conventional modeling techniques. Quantile regression (QR) not only offers a flexible way to capture heterogeneous effects across the conditional distribution of an outcome, but also naturally produces prediction intervals that help quantify uncertainty in future predictions. However, classical QR methods can face serious memory and computational constraints in large-scale settings. These limitations motivate the use of parallel computing to maintain tractability. While extensive work has examined sample-splitting strategies in settings where the number of observations $n$ greatly exceeds the number of features $p$, the equally important (ultra) high-dimensional regime ($p >> n$) has been comparatively underexplored. To address this gap, we introduce a feature-splitting proximal point algorithm, FS-QRPPA, for penalized QR in high-dimensional regime. Leveraging recent developments in variational analysis, we establish a Q-linear convergence rate for FS-QRPPA and demonstrate its superior scalability in large-scale genomic applications from the UK Biobank relative to existing methods. Moreover, FS-QRPPA yields more accurate coefficient estimates and better coverage for prediction intervals than current approaches. We provide a parallel implementation in the R package fsQRPPA, making penalized QR tractable on large-scale datasets.

Bayesian network models (Erdos Renyi, stochastic block models, random dot product graphs, graphons) are widely used in neuroscience, epidemiology, and the social sciences, yet real networks are sparse, heterogeneous, and exhibit higher-order dependence. How stable are network-based decisions, model selection, and policy recommendations to small model misspecification? We study local decision-theoretic robustness by allowing the posterior to vary within a small Kullback-Leibler neighborhood and choosing actions that minimize worst-case posterior expected loss. Exploiting low-dimensional functionals available under exchangeability, we (i) adapt decision-theoretic robustness to exchangeable graphs via graphon limits and derive sharp small-radius expansions of robust posterior risk; under squared loss the leading inflation is controlled by the posterior variance of the loss, and for robustness indices that diverge at percolation/fragmentation thresholds we obtain a universal critical exponent describing the explosion of decision uncertainty near criticality. (ii) Develop a nonparametric minimax theory for robust model selection between sparse Erdos-Renyi and block models, showing-via robustness error exponents-that no Bayesian or frequentist method can uniformly improve upon the decision-theoretic limits over configuration models and sparse graphon classes for percolation-type functionals. (iii) Propose a practical algorithm based on entropic tilting of posterior or variational samples, and demonstrate it on functional brain connectivity and Karnataka village social networks.

We introduce Conformal Interquantile Regression (CIR), a conformal regression method that efficiently constructs near-minimal prediction intervals with guaranteed coverage. CIR leverages black-box machine learning models to estimate outcome distributions through interquantile ranges, transforming these estimates into compact prediction intervals while achieving approximate conditional coverage. We further propose CIR+ (Conditional Interquantile Regression with More Comparison), which enhances CIR by incorporating a width-based selection rule for interquantile intervals. This refinement yields narrower prediction intervals while maintaining comparable coverage, though at the cost of slightly increased computational time. Both methods address key limitations of existing distributional conformal prediction approaches: they handle skewed distributions more effectively than Conformalized Quantile Regression, and they achieve substantially higher computational efficiency than Conformal Histogram Regression by eliminating the need for histogram construction. Extensive experiments on synthetic and real-world datasets demonstrate that our methods optimally balance predictive accuracy and computational efficiency compared to existing approaches.

Reporting test-retest reliability using the intraclass correlation coefficient (ICC) has received increasing attention due to the criticisms of poor transparency and replicability in neuroimaging research, as well as many other biomedical studies. Numerous studies have thus evaluated the reliability of their findings by comparing ICCs, however, they often failed to test statistical differences between ICCs or report confidence intervals. Relying solely on point estimates may preclude valid inference about population-level differences and compromise the reliability of conclusions. To address this issue, this study systematically reviewed the use of ICC in articles published in NeuroImage from 2022 to 2024, highlighting the prevalence of misreporting and misuse of ICCs. We further provide practical guidelines for conducting appropriate statistical inference on ICCs. For practitioners in this area, we introduce an online application for statistical testing and sample size estimation when utilizing ICCs. We recalculated confidence intervals and formally tested ICC values reported in the reviewed articles, thereby reassessing the original inferences. Our results demonstrate that exclusive reliance on point estimates could lead to unreliable or even misleading conclusions. Specifically, only two of the eleven reviewed articles provided unequivocally valid statistical inferences based on ICCs, whereas two articles failed to yield any valid inference at all, raising serious concerns about the replicability of findings in this field. These results underscore the urgent need for rigorous inferential frameworks when reporting and interpreting ICCs.

Language is a uniquely human trait, conveying information efficiently by organizing word sequences in sentences into hierarchical structures. A central question persists: Why is human language hierarchical? In this study, we show that hierarchization optimally solves the challenge of our limited working memory capacity. We established a likelihood function that quantifies how well the average number of units according to the language processing mechanisms aligns with human working memory capacity (WMC) in a direct fashion. The maximum likelihood estimate (MLE) of this function, tehta_MLE, turns out to be the mean of units. Through computational simulations of symbol sequences and validation analyses of natural language sentences, we uncover that compared to linear processing, hierarchical processing far surpasses it in constraining the tehta_MLE values under the human WMC limit, along with the increase of sequence/sentence length successfully. It also shows a converging pattern related to children's WMC development. These results suggest that constructing hierarchical structures optimizes the processing efficiency of sequential language input while staying within memory constraints, genuinely explaining the universal hierarchical nature of human language.

We study the problem of sampling from a target distribution $\pi(q)\propto e^{-U(q)}$ on $\mathbb{R}^d$, where $U$ can be non-convex, via the Hessian-free high-resolution (HFHR) dynamics, which is a second-order Langevin-type process that has $e^{-U(q)-\frac12|p|^2}$ as its unique invariant distribution, and it reduces to kinetic Langevin dynamics (KLD) as the resolution parameter $\alpha\to0$. The existing theory for HFHR dynamics in the literature is restricted to strongly-convex $U$, although numerical experiments are promising for non-convex settings as well. We focus on studying the convergence of HFHR dynamics when $U$ can be non-convex, which bridges a gap between theory and practice. Under a standard assumption of dissipativity and smoothness on $U$, we adopt the reflection/synchronous coupling method. This yields a Lyapunov-weighted Wasserstein distance in which the HFHR semigroup is exponentially contractive for all sufficiently small $\alpha>0$ whenever KLD is. We further show that, under an additional assumption that asymptotically $\nabla U$ has linear growth at infinity, the contraction rate for HFHR dynamics is strictly better than that of KLD, with an explicit gain. As a case study, we verify the assumptions and the resulting acceleration for three examples: a multi-well potential, Bayesian linear regression with $L^p$ regularizer and Bayesian binary classification. We conduct numerical experiments based on these examples, as well as an additional example of Bayesian logistic regression with real data processed by the neural networks, which illustrates the efficiency of the algorithms based on HFHR dynamics and verifies the acceleration and superior performance compared to KLD.

Feature selection is essential for high-dimensional biomedical data, enabling stronger predictive performance, reduced computational cost, and improved interpretability in precision medicine applications. Existing approaches face notable challenges. Filter methods are highly scalable but cannot capture complex relationships or eliminate redundancy. Deep learning-based approaches can model nonlinear patterns but often lack stability, interpretability, and efficiency at scale. Single-head attention improves interpretability but is limited in capturing multi-level dependencies and remains sensitive to initialization, reducing reproducibility. Most existing methods rarely combine statistical interpretability with the representational power of deep learning, particularly in ultra-high-dimensional settings. Here, we introduce MAFS (Multi-head Attention-based Feature Selection), a hybrid framework that integrates statistical priors with deep learning capabilities. MAFS begins with filter-based priors for stable initialization and guide learning. It then uses multi-head attention to examine features from multiple perspectives in parallel, capturing complex nonlinear relationships and interactions. Finally, a reordering module consolidates outputs across attention heads, resolving conflicts and minimizing information loss to generate robust and consistent feature rankings. This design combines statistical guidance with deep modeling capacity, yielding interpretable importance scores while maximizing retention of informative signals. Across simulated and real-world datasets, including cancer gene expression and Alzheimer's disease data, MAFS consistently achieves superior coverage and stability compared with existing filter-based and deep learning-based alternatives, offering a scalable, interpretable, and robust solution for feature selection in high-dimensional biomedical data.

Clustering is a central tool in biomedical research for discovering heterogeneous patient subpopulations, where group boundaries are often diffuse rather than sharply separated. Traditional methods produce hard partitions, whereas soft clustering methods such as fuzzy $c$-means (FCM) allow mixed memberships and better capture uncertainty and gradual transitions. Despite the widespread use of FCM, principled statistical inference for fuzzy clustering remains limited. We develop a new framework for weighted fuzzy $c$-means (WFCM) for settings with potential cluster size imbalance. Cluster-specific weights rebalance the classical FCM criterion so that smaller clusters are not overwhelmed by dominant groups, and the weighted objective induces a normalized density model with scale parameter $\sigma$ and fuzziness parameter $m$. Estimation is performed via a blockwise majorize–minimize (MM) procedure that alternates closed-form membership and centroid updates with likelihood-based updates of $(\sigma,\bw)$. The intractable normalizing constant is approximated by importance sampling using a data-adaptive Gaussian mixture proposal. We further provide likelihood ratio tests for comparing cluster centers and bootstrap-based confidence intervals. We establish consistency and asymptotic normality of the maximum likelihood estimator, validate the method through simulations, and illustrate it using single-cell RNA-seq and Alzheimer disease Neuroimaging Initiative (ADNI) data. These applications demonstrate stable uncertainty quantification and biologically meaningful soft memberships, ranging from well-separated cell populations under imbalance to a graded AD versus non-AD continuum consistent with disease progression.

We propose the first approach for multiple multivariate density-density regression (MDDR), making it possible to consider the regression of a multivariate density-valued response on multiple multivariate density-valued predictors. The core idea is to define a fitted distribution using a sliced Wasserstein barycenter (SWB) of push-forwards of the predictors and to quantify deviations from the observed response using the sliced Wasserstein (SW) distance. Regression functions, which map predictors' supports to the response support, and barycenter weights are inferred within a generalized Bayes framework, enabling principled uncertainty quantification without requiring a fully specified likelihood. The inference process can be seen as an instance of an inverse SWB problem. We establish theoretical guarantees, including the stability of the SWB under perturbations of marginals and barycenter weights, sample complexity of the generalized likelihood, and posterior consistency. For practical inference, we introduce a differentiable approximation of the SWB and a smooth reparameterization to handle the simplex constraint on barycenter weights, allowing efficient gradient-based MCMC sampling. We demonstrate MDDR in an application to inference for population-scale single-cell data. Posterior analysis under the MDDR model in this example includes inference on communication between multiple source/sender cell types and a target/receiver cell type. The proposed approach provides accurate fits, reliable predictions, and interpretable posterior estimates of barycenter weights, which can be used to construct sparse cell-cell communication networks.

Novelty detection via conformal $p$-values and BH procedure provides distribution-free global false discovery rate (FDR) control. We present here fundamental limits of this approach by showing that it does not produce reliable detection at the decision boundary. We study boundary false discovery rate (bFDR), the probability that the least extreme reported novelty is in fact a null observation. We first show that the support line (SL) procedure, controlling the bFDR in the continuous independent framework, fails to control the bFDR in the conformal case. We then present several modifications of the SL procedure that restore reliability at the decision boundary, by controlling the bFDR, with specific improvements in situations where many novelties are expected (adaptive procedures) and when the calibration sample is too small with respect to the test sample (subsampled procedures). Numerical experiments with both synthetic and real data support our findings and show the relevance of the new proposed approach.

Large language model fine-tuning is bottlenecked by memory: a 7B parameter model requires 84GB–14GB for weights, 14GB for gradients, and 56GB for FP32 optimizer states–exceeding even A100-40GB capacity. We present Chronicals, an open-source training framework achieving 3.51x speedup over Unsloth through four synergistic optimizations: (1) fused Triton kernels eliminating 75% of memory traffic via RMSNorm (7x), SwiGLU (5x), and QK-RoPE (2.3x) fusion; (2) Cut Cross-Entropy reducing logit memory from 5GB to 135MB through online softmax computation; (3) LoRA+ with theoretically-derived 16x differential learning rates between adapter matrices; and (4) Best-Fit Decreasing sequence packing recovering 60-75% of compute wasted on padding. On Qwen2.5-0.5B with A100-40GB, Chronicals achieves 41,184 tokens/second for full fine-tuning versus Unsloth's 11,736 tokens/second (3.51x). For LoRA at rank 32, we reach 11,699 tokens/second versus Unsloth MAX's 2,857 tokens/second (4.10x). Critically, we discovered that Unsloth's reported 46,000 tokens/second benchmark exhibited zero gradient norms–the model was not training. We provide complete mathematical foundations: online softmax correctness proofs, FlashAttention IO complexity bounds O(N^2 d^2 M^{-1}), LoRA+ learning rate derivations from gradient magnitude analysis, and bin-packing approximation guarantees. All implementations, benchmarks, and proofs are available at https://github.com/Ajwebdevs/Chronicals with pip installation via https://pypi.org/project/chronicals/.

Recent work has demonstrated surprisingly good performance of pre-trained LLMs on regression tasks (for example, time-series prediction), with the ability to incorporate expert prior knowledge and the information contained in textual metadata. However we observe major error cascades even in short sequences < ~100 points; these models are also computationally intensive and difficult to parallelise. Marginal LLM predictions do not suffer this issue and are trivially parallelised, but can predict over-broad densities. To address this, we propose combining these densities with a lightweight (diffusion-based) neural process. We show that this combination leads to better-calibrated predictions overall, outputs locally consistent trajectories, and leads to text-conditioned function space selection in the meta-learner. As part of this work we propose a gradient-free (and non-Monte Carlo) method for sampling from a product-of-experts of a score model and an 'expert' (here the LLM predictive densities). We believe this general method is of independent interest as it is applicable whenever an expert can be convolved with a Gaussian in closed form.

Cardiovascular disease (CVD) continues to be the major cause of death globally, calling for predictive models that not only handle diverse and high-dimensional biomedical signals but also maintain interpretability and privacy. We create a single multimodal learning framework that integrates cross modal transformers with graph neural networks and causal representation learning to measure personalized CVD risk. The model combines genomic variation, cardiac MRI, ECG waveforms, wearable streams, and structured EHR data to predict risk while also implementing causal invariance constraints across different clinical subpopulations. To maintain transparency, we employ SHAP based feature attribution, counterfactual explanations and causal latent alignment for understandable risk factors. Besides, we position the design in a federated, privacy, preserving optimization protocol and establish rules for convergence, calibration and uncertainty quantification under distributional shift. Experimental studies based on large-scale biobank and multi institutional datasets reveal state discrimination and robustness, exhibiting fair performance across demographic strata and clinically distinct cohorts. This study paves the way for a principled approach to clinically trustworthy, interpretable and privacy respecting CVD prediction at the population level.

Genome-wide association studies (GWAS) often find association signals between many genetic variants and traits of interest in a genomic region. Functional annotations of these variants provide valuable prior information that helps prioritize biologically relevant variants and enhances the power to detect causal variants. However, due to substantial correlations among these variants, a critical question is how to rigorously control the false discovery rate while effectively leveraging prior knowledge. We introduce annotation-informed knockoffs (AnnoKn), a knockoff-based method that performs annotation-informed variable selection with strict control of the false discovery rate. AnnoKn integrates the knockoff procedure with adaptive Lasso regression to evaluate the importance of multiple covariates while incorporating functional annotation information within a unified Bayesian framework. To facilitate real-world applications where individual-level data are not accessible, we further extend AnnoKn to operate on summary statistics. Through simulations and real-world applications to GTEx and GWAS datasets, we show that AnnoKn achieves superior power in detecting causal genetic variants compared with existing annotation-informed variable selection methods, while maintaining valid control over false discoveries.

We propose a novel finite-sample procedure for testing composite null hypotheses. Traditional likelihood ratio tests based on asymptotic $\chi^2$ approximations can exhibit substantial size distortions in small samples. Our procedure rejects the composite null hypothesis $H_0: \theta \in \Theta_0$ if the simple null hypothesis $H_{0,t}: \theta = \theta_t$ is rejected for every $\theta_t \in \Theta_0$, using a suitably inflated significance level. We derive a general formula for calibrating this inflated level based on the geometry of the null region $\Theta_0$. By employing finite-sample methods for testing the simple hypotheses $H_{0,t}: \theta = \theta_t$, the proposed pointwise testing procedure reduces the discrepancy between the actual and nominal significance levels, even in small samples. Whereas the traditional likelihood ratio test applies when the null region is defined solely by equality constraints, the proposed approach extends to null hypotheses defined by both equality and inequality constraints. We also discuss the application of the pointwise testing framework to composite null hypotheses expressed as unions of several component regions and to models involving nuisance parameters. Through several examples, we demonstrate numerically that the proposed test achieves accurate Type I error control in both small- and large-sample settings.

Artificial intelligence has advanced rapidly through large neural networks trained on massive datasets using thousands of GPUs or TPUs. Such training can occupy entire data centers for weeks and requires enormous computational and energy resources. Yet the optimization algorithms behind these runs have not kept pace. Most large scale training still relies on synchronous methods, where workers must wait for the slowest device, wasting compute and amplifying the effects of hardware and network variability. Removing synchronization seems like a simple fix, but asynchrony introduces staleness, meaning updates computed on outdated models. This makes analysis difficult, especially when delays arise from system level randomness rather than algorithmic choices. As a result, the time complexity of asynchronous methods remains poorly understood. This dissertation develops a rigorous framework for asynchronous first order stochastic optimization, focusing on the core challenge of heterogeneous worker speeds. Within this framework, we show that with proper design, asynchronous SGD can achieve optimal time complexity, matching guarantees previously known only for synchronous methods. Our first contribution, Ringmaster ASGD, attains optimal time complexity in the homogeneous data setting by selectively discarding stale updates. The second, Ringleader ASGD, extends optimality to heterogeneous data, common in federated learning, using a structured gradient table mechanism. Finally, ATA improves resource efficiency by learning worker compute time distributions and allocating tasks adaptively, achieving near optimal wall clock time with less computation. Together, these results establish asynchronous optimization as a theoretically sound and practically efficient foundation for distributed learning, showing that coordination without synchronization can be both feasible and optimal.

Anomaly detection is crucial in industrial applications for identifying rare and unseen patterns to ensure system reliability. Traditional models, trained on a single class of normal data, struggle with real-world distributions where normal data exhibit diverse patterns, leading to class imbalance and long-tailed anomaly score distributions (LTD). This imbalance skews model training and degrades detection performance, especially for minority instances. To address this issue, we propose a novel importance-weighted loss designed specifically for anomaly detection. Compared to the previous method for LTD in classification, our method does not require prior knowledge of normal data classes. Instead, we introduce a weighted loss function that incorporates importance sampling to align the distribution of anomaly scores with a target Gaussian, ensuring a balanced representation of normal data. Extensive experiments on three benchmark image datasets and three real-world hyperspectral imaging datasets demonstrate the robustness of our approach in mitigating LTD-induced bias. Our method improves anomaly detection performance by 0.043, highlighting its effectiveness in real-world applications.

A common approach to out-of-distribution prediction restricts models to causal or invariant covariates to avoid spurious associations that may change across environments. Despite its theoretical appeal, this strategy can underperform empirical risk minimization when only a subset of the causal parents of the outcome is observed. In such settings, non-causal covariates can serve as proxies for unobserved causal parents and improve prediction when the proxy relationship is stable, but they can hurt when shifts disrupt that relationship. Thus, the optimal covariate set can depend on the specific shift encountered. Because different shifts leave signatures in the unlabeled covariate distribution, we propose an environment-adaptive covariate selection algorithm that maps environment-level summaries to environment-specific covariate sets. These summaries may be hand-crafted or learned from multi-environment data, and prior causal knowledge can be incorporated as constraints. Across simulations and applied datasets, the proposed method improves over static causal, invariant, and other non-adaptive rules under diverse shifts.

Statistical inference for spatial processes from partially realized or scattered data has seen voluminous developments in diverse areas ranging from environmental sciences to business and economics. Inference on the associated rates of change has seen some recent developments. The literature has been restricted to Euclidean domains, where inference is sought on directional derivatives, rates along a chosen direction of interest, at arbitrary locations. Inference for higher order rates, particularly directional curvature has also proved useful in these settings. Modern spatial data often arise from non-Euclidean domains. This manuscript particularly considers spatial processes defined over compact Riemannian manifolds. We develop a comprehensive inferential framework for spatial rates of change for such processes over vector fields. In doing so, we formalize smoothness of process realizations and construct differential processes – the derivative and curvature processes. We derive conditions for kernels that ensure the existence of these processes and establish validity of the joint multivariate process consisting of the “parent” Gaussian process (GP) over the manifold and the associated differential processes. Predictive inference on these rates is devised conditioned on the realized process over the manifold. Manifolds arise as polyhedral meshes in practice. The success of our simulation experiments for assessing derivatives for processes observed over such meshes validate our theoretical findings. By enhancing our understanding of GPs on manifolds, this manuscript unlocks a variety of potential applications in machine learning and statistics where GPs have seen wide usage. We propose a fully model-based approach to inference on the differential processes arising from a spatial process from partially observed or realized data across scattered location on a manifold.

Estimating how different brain regions communicate with each other using EEG data is valuable both for medical research and clinical diagnosis. This involves quantifying the statistical dependencies among the activities of different brain areas, captured by the time-varying electric field recorded by scalp sensors. These dependencies can vary within and across individuals also in relationship with external factors such as age, mental state, or disease severity. Motivated by this problem, we propose a novel neighbor selection approach based on Gaussian functional graphical models and functional-on-functional regression to identify which brain regions interact and how interaction strength changes with individual features or covariates (e.g., age or clinical status). Our approach is fully automated and data-driven, and, in principle, can handle any number of continuous and categorical covariates simultaneously. Unlike existing approaches, it also produces results that are easy to interpret: one can directly assess whether the strength of each estimated interaction increases or decreases as the value of a given covariate varies. We evaluate our method through extensive simulation experiments and an application to real EEG data. The results demonstrate clear advantages over existing approaches, including more accurate estimation of brain connections and reduced computational cost, especially in high-dimensional settings involving a large number of brain regions and large sample sizes.

Effective data center cooling is crucial for reliable operation; however, cooling systems often exhibit inefficiencies that result in excessive energy consumption. This paper presents a three-stage, physics-guided machine learning framework for identifying and reducing cooling energy waste in high-performance computing facilities. Using one year of 10-minute resolution operational data from the Frontier exascale supercomputer, we first train a monotonicity-constrained gradient boosting surrogate that predicts facility accessory power from coolant flow rates, temperatures, and server power. The surrogate achieves a mean absolute error of 0.026 MW and predicts power usage effectiveness within 0.01 of measured values for 98.7% of test samples. In the second stage, the surrogate serves as a physics-consistent baseline to quantify excess cooling energy, revealing approximately 85 MWh of annual inefficiency concentrated in specific months, hours, and operating regimes. The third stage evaluates guardrail-constrained counterfactual adjustments to supply temperature and subloop flows, demonstrating that up to 96% of identified excess can be recovered through small, safe setpoint changes while respecting thermal limits and operational constraints. The framework yields interpretable recommendations, supports counterfactual analyses such as flow reduction during low-load periods and redistribution of thermal duty across cooling loops, and provides a practical pathway toward quantifiable reductions in accessory power. The developed framework is readily compatible with model predictive control and provides a template that, with site-specific recalibration, could be adapted to other liquid-cooled data centers with different configurations and cooling requirements.

We investigate convergence of the expectation maximization algorithm by representing it as a generalized proximal method. Convergence of iterates and not just in value is investigated under natural hypotheses such as definability of the incomplete data log-likelihood in the sense of o-minimal structure theory.

Graphs arise across diverse domains, from biology and chemistry to social and information networks, as well as in transportation and logistics. Inference on graph-structured data requires methods that are permutation-invariant, scalable across varying sizes and sparsities, and capable of capturing complex long-range dependencies, making posterior estimation on graph parameters particularly challenging. Amortized Bayesian Inference (ABI) is a simulation-based framework that employs generative neural networks to enable fast, likelihood-free posterior inference. We adapt ABI to graph data to address these challenges to perform inference on node-, edge-, and graph-level parameters. Our approach couples permutation-invariant graph encoders with flexible neural posterior estimators in a two-module pipeline: a summary network maps attributed graphs to fixed-length representations, and an inference network approximates the posterior over parameters. In this setting, several neural architectures can serve as the summary network. In this work we evaluate multiple architectures and assess their performance on controlled synthetic settings and two real-world domains - biology and logistics - in terms of recovery and calibration.

This study presents an Initial Data Analysis (IDA) of the German Transplantation Registry (TxReg) data for a better data understanding and to inform future data analyses. The IDA is focusing on data on first-time kidney-only transplantations in adult recipients from deceased donors between 2006 and 2016 and refers to data from 14,954 recipients and 9,964 donors across 25 tables. Investigated aspects include missing data patterns and structure, data consistency, and availability of event time data. Results show that missing data proportions vary widely, with some tables nearly complete while others have over 50% missing values. Missing data patterns are identified using a decision tree approach. An influx and outflux analysis demonstrates that some variables have high potential for imputing missing data, while others were less suitable for imputation. We identified 168 multi-sourced variables that are reported by multiple data providers in parallel leading to discrepancies for some variables but also providing opportunities for missing data imputation. Our findings on event time data demonstrate the importance of carefully selecting the variables used for event time analyses as results will strongly depend on this selection. In summary, our findings highlight the challenges when utilizing the TxReg data for research and provide recommendations for data preprocessing and analysis in future analyses.

We study the extent to which standard machine learning algorithms rely on exchangeability and independence of data by introducing a monotone adversarial corruption model. In this model, an adversary, upon looking at a "clean" i.i.d. dataset, inserts additional "corrupted" points of their choice into the dataset. These added points are constrained to be monotone corruptions, in that they get labeled according to the ground-truth target function. Perhaps surprisingly, we demonstrate that in this setting, all known optimal learning algorithms for binary classification can be made to achieve suboptimal expected error on a new independent test point drawn from the same distribution as the clean dataset. On the other hand, we show that uniform convergence-based algorithms do not degrade in their guarantees. Our results showcase how optimal learning algorithms break down in the face of seemingly helpful monotone corruptions, exposing their overreliance on exchangeability.

We propose a graph-based, multi-fidelity learning framework for the prediction of stationary Navier–Stokes solutions in non-parametrized two-dimensional geometries. The method is designed to guide the learning process through successive approximations, starting from reduced-order and full Stokes models, and progressively approaching the Navier–Stokes solution. To effectively capture both local and long-range dependencies in the velocity and pressure fields, we combine graph neural networks with Transformer and Mamba architectures. While Transformers achieve the highest accuracy, we show that Mamba can be successfully adapted to graph-structured data through an unsupervised node-ordering strategy. The Mamba approach significantly reduces computational cost while maintaining performance. Physical knowledge is embedded directly into the architecture through an encoding-processing-physics informed decoding pipeline. Derivatives are computed through algebraic operators constructed via the Weighted Least Squares method. The flexibility of these operators allows us not only to make the output obey the governing equations, but also to constrain selected hidden features to satisfy mass conservation. We introduce additional physical biases through an enriched graph convolution with the same differential operators describing the PDEs. Overall, we successfully guide the learning process by physical knowledge and fluid dynamics insights, leading to more regular and accurate predictions

Curve registration plays a major role in functional data analysis by separating amplitude and phase variation through warping functions and the accurate simulation of warping processes is essential for developing statistical methods that properly account for phase variability in functional data. In this paper, we focus on the simulation of continuous warping processes with a prescribed expectation and a controllable variance. We study and compare three procedures, including two existing methods and a new algorithm based on randomized empirical cumulative distribution functions. For each approach, we provide an operational description and establish theoretical results for the first two moments of the simulated processes. A numerical study illustrates the theoretical findings and highlights the respective merits of the three methods. Finally, we present an application to the analysis of hourly temperature data spatially averaged over Canada, based on simulated realizations from a warping process estimated using spatially averaged temperature quantile functions.

We study the asymptotic behaviour of modified weighted power variations of the Hermite process of arbitrary order. By selecting suitable "good" increments and exploiting their decomposition into dominant independent components, we establish a central limit theorem for weighted $p$-variations using tools from Stein-Malliavin calculus. Our results extend previous works on modified quadratic and wavelet-based variations to general powers and to weighted settings, with explicit bounds in Wasserstein distance. We further apply these limit theorems to construct asymptotically Gaussian estimators of integrated volatility in Hermite-driven models, thereby extending fBm-based methods to non-Gaussian settings. The last part of our work contains numerical simulations which illustrate the practical performance of the proposed estimators.

Arrhythmia is an abnormality of the heart's rhythm, caused by problems in the conductive system and resulting in irregular heartbeats. There is increasing evidence that undertaking frequent endurance sports training elevates one's risk of arrhythmia. Arrhythmia is diagnosed using an electrocardiogram (ECG) but this is not typically available to athletes while exercising. Previous research by Crickles investigates the usefulness of commonly available heart rate data in detecting signs of arrhythmia. It is hypothesised that a feature termed 'gappiness', defined by jumps in the heart rate while the athlete is under exertion, may be a characteristic of arrhythmia. A correlation was found between the proportion of 'gappy' activities and survey responses about heart rhythm problems. We develop on this measure by exploring various methods to detect spikes in heart rate data, allowing us to describe the extent of irregularity in an activity via the rate of spikes. We first compare the performance of these methods on simulated data, where we find that smoothing using a moving average and setting a constant threshold on the residuals is most effective. This method was then implemented on real data provided by Crickles from 168 athletes, where no significant correlation was found between the spike rates and survey responses. However, when considering only those spikes that occur above a heart rate of 160 beats per minute (bpm) a significant correlation was found. This supports the hypothesis that jumps at only high heart rates are informative of arrhythmia and indicates the need for further research into better measures to characterise features of heart rate data.

This study evaluates the performance of various classifiers in three distinct models: response, risk, and response-risk, concerning credit card mail campaigns and default prediction. In the response model, the Extra Trees classifier demonstrates the highest recall level (79.1%), emphasizing its effectiveness in identifying potential responders to targeted credit card offers. Conversely, in the risk model, the Random Forest classifier exhibits remarkable specificity of 84.1%, crucial for identifying customers least likely to default. Furthermore, in the multi-class response-risk model, the Random Forest classifier achieves the highest accuracy (83.2%), indicating its efficacy in discerning both potential responders to credit card mail campaign and low-risk credit card users. In this study, we optimized various performance metrics to solve a specific credit risk and mail responsiveness business problem.