arXiv subjects

Xuefei Lu

Publications and source records attributed to Xuefei Lu.

6 recordsLinked to original sources

arXiv2026

Smart predict-then-robustly-optimize

In this paper, we propose and study a robust variant of the smart predict-then-optimize approach that accounts for prediction shifts due to disturbance in the covariate feature space. While traditional integrated-learning-and-optimization models assume that side information is perfectly revealed, empirical data-driven features are frequently corrupted or noisy at the time of decision-making, leading to fragile operational policies. To bridge this gap, we integrate principles of robust optimization directly into the predictive-prescriptive pipeline via a smart predict-then-robustly optimize loss and establish a computationally tractable convex surrogate, designed to hedge against worst-case feature perturbations. On the theoretical front, we formalize the structural validity of this surrogate by proving its approximation error probability decays exponentially according to a sub-Gaussian concentration profile. Furthermore, we establish that under mild assumptions, the surrogate is Fisher consistent with high probability. We also prove necessary conditions under which our framework outperforms standard smart predict-then-optimize and maintain its superiority even when the standard method is equipped with regularized upstream predictions. Numerical experiments validate that our robust framework consistently yields significant performance improvements over standard methods, both in out-of-sample terms and in training stability.

cs.LG↗

arXiv2026

Feature Attribution-Based Explainability Analysis of Deep Learning Models in Predictive Process Monitoring

Predictive process monitoring supports the optimization and control of operational business processes by forecasting the future state or outcome of ongoing cases. While deep neural networks have achieved strong performance for these tasks by modeling sequential dependencies in event logs, their black-box nature limits trust and practical adoption. Feature attribution methods are often used to address this, but applying them directly poses a dilemma: event-level attributions impose high computational complexity for long traces, while explanations based on aggregated trace representations often fail to capture the underlying control-flow dynamics. To address this issue, we propose a local post-hoc explainability method for deep neural networks in outcome prediction. The method relies on a control-flow-aware segmentation algorithm that partitions a trace into meaningful segments and supports the computation of segment-level SHAP explanations. This makes it possible to identify which parts of a trace influence a prediction and which change points steer the case toward the predicted outcome. We assess the proposed segmentation method on a synthetic dataset with known process logic, where meaningful change points can be explicitly verified, and we demonstrate its usefulness on real-world event logs from a loan application process and an administrative process of a Dutch municipality.

cs.LG↗

arXiv2026

Trustworthy Feature Importance Avoids Unrestricted Permutations

Feature importance methods using unrestricted permutations are flawed due to extrapolation errors; such errors appear in all non-trivial variable importance approaches. We propose three new approaches: conditional model reliance and Knockoffs with Gaussian transformation, and restricted ALE plot designs. Theoretical and numerical results show our strategies reduce/eliminate extrapolation.

stat.ML↗

arXiv2024

Dispersion Relations for Active Undulators in Overdamped Environments

Organisms that locomote by propagating waves of body bending can maintain performance across heterogeneous environments by modifying their gait frequency $\omega$ or wavenumber $k$. We identify a unifying relationship between these parameters for overdamped undulatory swimmers (including nematodes, spermatozoa, and mm-scale fish) moving in diverse environmental rheologies, in the form of an active `dispersion relation' $\omega\propto k^{\pm2}$. A model treating the organisms as actively driven viscoelastic beams reproduces the experimentally observed scaling. The relative strength of rate-dependent dissipation in the body and the environment determines whether $k^2$ or $k^{-2}$ scaling is observed. The existence of these scaling regimes reflects the $k$ and $\omega$ dependence of the various underlying force terms and how their relative importance changes with the external environment and the neuronally commanded gait.

physics.bio-ph↗

arXiv2020

Is Time to Intervention in the COVID-19 Outbreak Really Important? A Global Sensitivity Analysis Approach

Italy has been one of the first countries timewise strongly impacted by the COVID-19 pandemic. The adoption of social distancing and heavy lockdown measures is posing a heavy burden on the population and the economy. The timing of the measures has crucial policy-making implications. Using publicly available data for the pandemic progression in Italy, we quantitatively assess the effect of the intervention time on the pandemic expansion, with a methodology that combines a generalized susceptible-exposed-infectious-recovered (SEIR) model together with statistical learning methods. The modeling shows that the lockdown has strongly deviated the pandemic trajectory in Italy. However, the difference between the forecasts and real data up to 20 April 2020 can be explained only by the existence of a time lag between the actual issuance date and the full effect of the measures. To understand the relative importance of intervention with respect to other factors, a thorough uncertainty quantification of the model predictions is performed. Global sensitivity indices show that the the time of intervention is 4 times more relevant than quarantine, and eight times more important than intrinsic features of the pandemic such as protection and infection rates. The relevance of their interactions is also quantified and studied.

stat.AP↗

arXiv2019

Bayesian estimation of probabilistic sensitivity measures

Computer experiments are becoming increasingly important in scientific investigations. In the presence of uncertainty, analysts employ probabilistic sensitivity methods to identify the key-drivers of change in the quantities of interest. Simulation complexity, large dimensionality and long running times may force analysts to make statistical inference at small sample sizes. Methods designed to estimate probabilistic sensitivity measures at relatively low computational costs are attracting increasing interest. We propose a fully Bayesian approach to the estimation of probabilistic sensitivity measures based on a one-sample design. We discuss, first, new estimators based on placing piecewise constant priors on the conditional distributions of the output given each input, by partitioning the input space. We then present two alternatives, based on Bayesian non-parametric density estimation, which bypass the need for predefined partitions. In all cases, the Bayesian paradigm guarantees the quantification of uncertainty in the estimation process through the posterior distribution over the sensitivity measures, without requiring additional simulator evaluations. The performance of the proposed methods is compared to that of traditional point estimators in a series of numerical experiments comprising synthetic but challenging simulators, as well as a realistic application. $\textit{An Updated Version of the Manuscript is Forthcoming in Statistics and Computing.}$

stat.ME↗