Selected work

Representative Papers

Efficient Robust Conformal Prediction via Lipschitz-Bounded Networks

Jun 05, 2025

Traditional conformal prediction (CP) fails under adversarial attacks, while existing robust CP methods suffer from excessively large prediction sets or high computational overhead on large-scale tasks. To address this, we propose Lip-RCP—the first efficient robust prediction framework that deeply integrates 1-Lipschitz robust neural networks with CP. Methodologically, we impose Lipschitz constraints to ensure output stability and derive, for the first time, a theoretical worst-case coverage bound for standard CP under arbitrary attack magnitudes. Experiments on medium- and large-scale benchmarks (e.g., ImageNet) show that Lip-RCP reduces robust prediction set size by up to 42% over state-of-the-art methods while accelerating inference by 3.8×. Crucially, it strictly guarantees both nominal coverage ≥90% and finite-sample robust coverage—without compromising statistical validity.

1 citationsRead paper

A Model for Imbalanced Label Aggregation: A Focus on Minority-Class Detection

Jul 27, 2026

This work addresses the challenge of accurately recovering minority-class labels in crowdsourced annotation tasks under severe class imbalance. The authors propose a generative label aggregation model that jointly captures category-dependent item difficulty and annotator competence, departing from conventional assumptions in the field. By re-examining the Condorcet jury theorem under class-imbalanced settings, they theoretically demonstrate that majority voting asymptotically preserves the original class distribution. Empirical evaluation across 33 real-world multiclass crowdsourcing datasets shows that the proposed model substantially improves recall for minority classes while maintaining competitive overall accuracy.

0 citationsRead paper

LiST: Lipschitz Scaling Training for Robust and Calibrated Neural Networks

Jul 08, 2026

Existing methods struggle to simultaneously achieve high accuracy, robustness, and calibration in neural networks. This work proposes Lipschitz Scaling Training (LiST), which establishes, for the first time, a theoretical connection between Lipschitz constraints and temperature scaling. By dynamically adjusting the global Lipschitz constant during training, LiST embeds calibration directly into the learning process, automatically identifying a calibration-optimal operating point along the accuracy–robustness Pareto frontier. The method integrates margin-aware Lipschitz constraints, dynamic constant adaptation, and calibration-aware optimization, and further improves sample efficiency by reusing calibration data after convergence. Experiments on CIFAR-10, CIFAR-100, and Tiny-ImageNet demonstrate that LiST matches baseline performance in both accuracy and robustness while achieving well-calibrated predictions without any post-hoc processing.

0 citationsRead paper

LipSSD: Lipschitz-Constrained Single-Shot Detection for Adversarially Robust Object Detection

Jul 06, 2026

Existing object detectors exhibit insufficient robustness under adversarial attacks, limiting their applicability in safety-critical scenarios. This work proposes LipSSD, the first single-stage detector that systematically incorporates Lipschitz constraints to achieve intrinsic robustness. By redesigning the SSD architecture with Lipschitz-bounded components, LipSSD attains attack-agnostic defense capabilities without relying on adversarial training. Extensive experiments demonstrate that LipSSD significantly outperforms conventional adversarial training methods across multiple white-box attacks on Pascal VOC, LARD, and KITTI datasets. Notably, it improves mAP@50 by up to 15 percentage points against unseen attacks while maintaining competitive performance on clean samples.

0 citationsRead paper

RIDE: An Open Dataset and Benchmark for Train Delay Prediction

Jun 03, 2026

Progress in train delay prediction has been hindered by the absence of standardized datasets, task formulations, and evaluation protocols. This work proposes RIDE—the first open benchmark encompassing the entire Belgian railway network—integrating 94.5 million train events, 3.6 million journeys, and 35.7 million weather records from 2023 to 2025. RIDE establishes a hierarchical pipeline that transforms raw data into model-ready inputs and unifies prediction tasks with consistent evaluation metrics. It enables fair, multi-model comparisons and fine-grained performance analysis through relational modeling, graph neural networks, statistical methods, and deep learning approaches. Experimental results demonstrate that learning-based methods substantially outperform non-learning baselines, with graph neural networks achieving overall superior performance, as quantified by MAE, RMSE, and multidimensional performance breakdowns.

0 citationsRead paper

Recent publications

Latest Papers

A Model for Imbalanced Label Aggregation: A Focus on Minority-Class Detection

Jul 27, 2026

This work addresses the challenge of accurately recovering minority-class labels in crowdsourced annotation tasks under severe class imbalance. The authors propose a generative label aggregation model that jointly captures category-dependent item difficulty and annotator competence, departing from conventional assumptions in the field. By re-examining the Condorcet jury theorem under class-imbalanced settings, they theoretically demonstrate that majority voting asymptotically preserves the original class distribution. Empirical evaluation across 33 real-world multiclass crowdsourcing datasets shows that the proposed model substantially improves recall for minority classes while maintaining competitive overall accuracy.

0 citationsRead paper

LiST: Lipschitz Scaling Training for Robust and Calibrated Neural Networks

Jul 08, 2026

Existing methods struggle to simultaneously achieve high accuracy, robustness, and calibration in neural networks. This work proposes Lipschitz Scaling Training (LiST), which establishes, for the first time, a theoretical connection between Lipschitz constraints and temperature scaling. By dynamically adjusting the global Lipschitz constant during training, LiST embeds calibration directly into the learning process, automatically identifying a calibration-optimal operating point along the accuracy–robustness Pareto frontier. The method integrates margin-aware Lipschitz constraints, dynamic constant adaptation, and calibration-aware optimization, and further improves sample efficiency by reusing calibration data after convergence. Experiments on CIFAR-10, CIFAR-100, and Tiny-ImageNet demonstrate that LiST matches baseline performance in both accuracy and robustness while achieving well-calibrated predictions without any post-hoc processing.

0 citationsRead paper

LipSSD: Lipschitz-Constrained Single-Shot Detection for Adversarially Robust Object Detection

Jul 06, 2026

Existing object detectors exhibit insufficient robustness under adversarial attacks, limiting their applicability in safety-critical scenarios. This work proposes LipSSD, the first single-stage detector that systematically incorporates Lipschitz constraints to achieve intrinsic robustness. By redesigning the SSD architecture with Lipschitz-bounded components, LipSSD attains attack-agnostic defense capabilities without relying on adversarial training. Extensive experiments demonstrate that LipSSD significantly outperforms conventional adversarial training methods across multiple white-box attacks on Pascal VOC, LARD, and KITTI datasets. Notably, it improves mAP@50 by up to 15 percentage points against unseen attacks while maintaining competitive performance on clean samples.

0 citationsRead paper

RIDE: An Open Dataset and Benchmark for Train Delay Prediction

Jun 03, 2026

Progress in train delay prediction has been hindered by the absence of standardized datasets, task formulations, and evaluation protocols. This work proposes RIDE—the first open benchmark encompassing the entire Belgian railway network—integrating 94.5 million train events, 3.6 million journeys, and 35.7 million weather records from 2023 to 2025. RIDE establishes a hierarchical pipeline that transforms raw data into model-ready inputs and unifies prediction tasks with consistent evaluation metrics. It enables fair, multi-model comparisons and fine-grained performance analysis through relational modeling, graph neural networks, statistical methods, and deep learning approaches. Experimental results demonstrate that learning-based methods substantially outperform non-learning baselines, with graph neural networks achieving overall superior performance, as quantified by MAE, RMSE, and multidimensional performance breakdowns.

0 citationsRead paper

From SGD to Muon: Adaptive Optimization via Schatten-p Norms

May 19, 2026

This work addresses the limitation of existing optimizers, which rely on fixed geometric constraints that often fail to align with the intrinsic geometry of individual layers in deep neural networks. The authors propose a data-driven adaptive optimization framework that dynamically selects the optimal update geometry for each layer based on gradient and activation statistics, unifying optimizers such as SGD, Muon, and Adam as special cases. Their approach is the first to efficiently adapt the geometry of the linear minimization oracle (LMO) at runtime by integrating Schatten-p norms, a single-step stochastic feature regression surrogate model, and parameter-level preconditioning, incurring only approximately 3% additional computational overhead. Empirical evaluations across three training scenarios demonstrate that the method matches or surpasses the performance of state-of-the-art optimizers like Muon and AdamW, confirming the efficacy and scalability of adaptive geometric optimization.

0 citationsRead paper