cross-cohort calibration

Conventional covariate adjustment in randomized controlled trials (RCTs) often suffers from suboptimal efficiency due to inflexible, pre-specified modeling assumptions. Method: We propose a prespecified yet data-adaptive machine learning framework for RCT analysis. It introduces the first adaptive targeted maximum likelihood estimation (TMLE) strategy that integrates sample splitting with cross-validated variance minimization—enabling data-driven model selection while strictly adhering to the pre-specified statistical analysis plan (SAP). Rigorous validation employs adaptive pre-specification, plasmode simulation, and parametric simulation. Contribution/Results: Applied to primary endpoint analyses of eight published clinical trials (2022–2024), our method significantly improves precision in marginal treatment effect estimation. It supports real-time, global remote unblinding and robust implementation, establishing a new paradigm for RCT analysis that is prespecifiable, reproducible, and statistically efficient.

Existing post-hoc calibration methods for multi-class classifiers—particularly logistic regression–based approaches—suffer from overfitting due to excessive parameters and limited calibration data. Method: We propose Structured Matrix Scaling (SMS), a principled calibration framework built upon multinomial logistic regression, incorporating structured matrix regularization (e.g., low-rank or diagonal-plus-low-rank constraints), cross-class parameter sharing, and robust feature preprocessing. This design simultaneously enhances expressivity and controls variance. Contribution/Results: SMS theoretically overcomes the representational limitations of temperature scaling and standard matrix scaling, achieving a superior bias–variance trade-off. Extensive experiments across diverse models and datasets demonstrate that SMS significantly outperforms existing logistic regression–based calibration methods, while exhibiting strong scalability and practicality. The implementation is open-sourced, establishing a new efficient and robust benchmark for probabilistic calibration.

To address the challenge of online adaptation to dynamic data distributions after deployment of time-series foundation models, this paper proposes AdapTS, a lightweight online adaptation framework. Methodologically, AdapTS decouples foundation model inference from adapter learning via two novel modules: AdapTS-Forecaster—a linear or MLP-based lightweight time-series modeling component—and AdapTS-Weighter—a gradient-free, dynamically weighted fusion mechanism. It further incorporates online distribution estimation and real-time feedback-driven calibration. Crucially, AdapTS enables zero-shot fine-tuning and low-overhead adaptation without modifying the frozen foundation model. Evaluated across multiple standard benchmarks, AdapTS consistently improves the average MSE of mainstream time-series foundation models by 7.2%–15.8%, while incurring less than 3 ms additional inference latency. This demonstrates substantial gains in both predictive accuracy and practical deployability.

To address the challenge of balancing treatment effect heterogeneity learning and decision timeliness in adaptive clinical trials, this paper proposes a covariate-adjusted response-adaptive design that accelerates randomization probability updates using surrogate outcomes. Methodologically, it introduces the first formal framework quantifying the dual benefits—accelerated learning and bias reduction—of surrogate outcomes in sequential adaptive trials; develops an Online-Superlearner–driven mechanism for dynamic surrogate selection; and establishes a model-agnostic, targeted minimum loss–based estimation (TMLE) inference method tailored to adaptive trial data. Extensive simulations—including scenarios calibrated to real trial data—demonstrate that the design significantly improves the expected outcome for newly enrolled participants while preserving asymptotic normality of estimators. Collectively, it provides a generalizable toolkit for evaluation, selection, and inference in adaptive clinical trials.

This work addresses the challenge of verifying covariate balance in covariate shift adaptation by proposing a sequentially valid, anytime-stoppable validation framework. Built upon time-uniform confidence sequences, the method dynamically monitors covariate balance for a pre-specified function class within a prescribed tolerance band and terminates as soon as all target moments fall within this band, thereby certifying balance. Its key contribution lies in providing, for the first time, a locally and absolutely valid certification of balance for any adjustment strategy, supporting data-dependent stopping times while rigorously controlling the probability of erroneous balance confirmation. Integrated with KL-divergence-based drift diagnostics and detection of admissible adjustment regions, experiments demonstrate the method’s superior performance in error rate control, locality with respect to function classes, effectiveness in drift diagnosis, and guaranteed coverage in conformal prediction.

This study addresses a key challenge in randomized controlled trials: how to effectively leverage covariate adjustment to improve the precision of average treatment effect estimation while satisfying regulatory requirements and ensuring statistical validity. The authors propose a prespecified, transparent, and reproducible covariate adjustment framework that, for the first time, integrates data-adaptive methods and machine learning into a regulatory-compliant analytical pipeline. By combining model-misspecification-robust estimation with semiparametric efficiency theory, the approach consistently outperforms unadjusted analyses without compromising causal interpretability or statistical validity. It substantially enhances estimation precision, increases statistical power, and yields narrower confidence intervals.

This work addresses the challenge of enabling predictive systems to dynamically adapt their behavior based on contextual information for personalized inference. To this end, it proposes a unified framework that maps context into adaptation parameters for prediction and, for the first time, establishes a mathematical equivalence between explicit parameter adaptation and implicit expert routing under kernel ridge regression. The framework theoretically unifies diverse methodologies—including varying-coefficient models, local regression, prompt engineering, retrieval-augmented approaches, and mixture-of-experts—under fixed features and squared loss. Key contributions include deriving a general formulation for context-adaptive inference, proposing practical design principles and evaluation metrics such as adaptation efficiency and routing stability, and highlighting critical open problems concerning identifiability and robustness under distributional shifts.

This work addresses Bayesian optimal experimental design under computationally expensive models with limited design evaluations. It proposes an adaptive sequential elimination algorithm that significantly reduces the variance and computational cost of nested Monte Carlo estimators by reusing parameter samples, employing common random numbers, and applying Rao–Blackwellization. A bootstrap-based probabilistic comparison mechanism is integrated to iteratively eliminate inferior designs. The method achieves high reliability while drastically reducing the number of model evaluations, making it well-suited for large-scale engineering applications where computational efficiency and decision accuracy must be carefully balanced.