arXiv subjects

Chengyao Yu

Publications and source records attributed to Chengyao Yu.

4 recordsLinked to original sources

arXiv2026

Robust Conformalized Selection with Noisy Responses

Conformalized selection has been widely applied to select high-quality candidates from large datasets with rigorous uncertainty quantification, such as reliable labeling, drug discovery, and the alignment of large language models. Nevertheless, existing methods assume clean responses on calibration data, an assumption that rarely holds in practice. In this paper, we formulate the above tasks as selecting candidates with true predicted labels or with responses exceeding certain values. We demonstrate that existing conformal selection methods fail to control the false discovery rate (FDR) or suffer from severe power loss under contaminated calibration data. To that end, we propose Robust Conformalized Selection (RCS), a unified framework for selective classification with valid FDR control under general label contamination. The key insight of RCS lies in a novel statistical reduction: by separately conditioning on different classes, we translate the intractable label noise into a localized covariate shift problem, which then enables a covariate-adjusted empirical-Bayes-type estimate of the number of false selections. Statistical properties such as the asymptotic FDR control, power optimality, and robustness of RCS are established. We further develop an instantiation of RCS under randomized response model, and also apply RCS to the task of selecting candidates with large response values. Extensive experiments on both simulated and real-world datasets demonstrate the effectiveness of RCS.

stat.ML↗

arXiv2026

Anytime Safe PAC Efficient Reasoning

Large Reasoning Models (LRMs) have demonstrated remarkable performance on complex tasks but suffer from high computational costs and latency. While selective thinking strategies improve efficiency by routing easy queries to non-thinking models, existing approaches often incur uncontrollable errors, especially in online settings where the performance loss of a non-thinking model is only partially observed and data are non-stationary. To address this, we propose Betting Probably Approximately Correct (B-PAC) reasoning, a principled method that enables anytime safe and efficient online reasoning under partial feedback. Specifically, we utilize inverse propensity scoring estimators to construct test supermartingales for candidate thresholds, and then dynamically adjust the routing threshold based on the accumulated statistical evidence of safety. Theoretically, we establish the anytime-valid performance loss control and the efficiency of B-PAC reasoning. Extensive experiments demonstrate that B-PAC reasoning significantly reduces computational overhead, decreasing thinking model usage by up to 81.01\%, while controlling the performance loss below the user-specified level.

cs.AI↗

arXiv2025

FAIR-Pruner: A Flexible Framework for Automatic Layer-Wise Pruning via Tolerance of Difference

Structured pruning is a standard tool for compressing deep neural networks, but its practical performance depends on how sparsity is allocated across layers. We propose FAIR-Pruner, a search-free framework for adaptive layer-wise structured pruning. FAIR-Pruner uses two within-layer rankings: a removal-oriented signal that proposes candidate units and a protection-oriented signal that identifies task-sensitive units. Its core component, Tolerance of Difference (ToD), measures the overlap between the removal prefix and the protected tail, and uses a shared tolerance level to induce non-uniform pruning depths across layers. As a default vision instantiation, FAIR-Pruner combines a Wasserstein-based U-Score for class-conditional unit separability with a Taylor-based R-Score for task-level sensitivity; the same ToD allocation rule can also be paired with alternative removal signals. Theoretically, we analyze ToD through the population R-Score, derive rank-based control of the high-R-Score mass entering the pruning set, and identify an additive exchange condition for same-budget comparison with uniform pruning. Experiments on CIFAR-10, CIFAR-100, SVHN, and ImageNet across VGG, ResNet, DenseNet, ConvNeXt, and DeiT show strong accuracy--compression trade-offs. Prune-only experiments on routed-expert Qwen1.5-MoE-A2.7B-Chat further examine architectural extensibility under matched expert budgets. FAIR-Pruner is released as a pip-installable open-source package.

cs.LG↗

arXiv2024

A generalized e-value feature detection method with FDR control at multiple resolutions

Multiple resolutions arise across a range of explanatory features due to domain-specific structures, leading to the formation of feature groups. It follows that the simultaneous detection of significant features and groups aimed at a specific response with false discovery rate (FDR) control stands as a crucial issue, such as the spatial genome-wide association studies. Nevertheless, existing detection methods with multilayer FDR control generally rely on valid p-values or knockoff statistics, which can be not flexible, powerful and stable in several settings. To fix this issue effectively, this article develops a novel method of Stabilized Flexible E-Filter Procedure (SFEFP), by constructing unified generalized e-values, leveraging a generalized e-filter, and adopting a stabilization treatment with power enhancement. This method flexibly incorporates diverse base detection procedures at different resolutions to provide consistent, powerful, and stable results, while controlling FDR at multiple resolutions simultaneously. Statistical properties of multilayer filtering procedure encompassing one-bit property, multilayer FDR control, and stability guarantee are established. We also develop several examples for SFEFP such as the eDS-filter. Simulation studies and the analysis of HIV mutation data demonstrate the efficacy of SFEFP.

stat.ME↗