Selected work

Representative Papers

Multi-Class vs. Multi-Label BERT for CVE-to-CWE Mapping: How Taxonomy Structure Shapes the Errors

Jul 08, 2026

This study addresses the ongoing challenge of automatically mapping Common Vulnerabilities and Exposures (CVEs) to Common Weakness Enumerations (CWEs), particularly the lack of systematic evaluation in choosing between multi-class and multi-label modeling strategies. Framing the task as text classification, the authors compare these two approaches using three transformer-based encoders—BERT Base, SecureBERT, and CySecBERT—across CWE label spaces of varying granularity. They further introduce a hierarchical relaxed evaluation metric to more accurately assess model performance within the structured hierarchy of CWEs. Experimental results show that the multi-class approach generally achieves higher macro F1 scores, though this advantage diminishes as the label space narrows. CySecBERT performs best under the multi-label setting. Notably, the hierarchical relaxed metric elevates macro F1 from approximately 81% to 90%, indicating that most misclassifications stem from ambiguities in the CWE taxonomy rather than model limitations.

0 citationsRead paper

Physics-Informed Machine Learning Under Small-Data Constraints: Lessons from Abrasive Waterjet Milling

Jul 08, 2026

This study addresses the challenges of applying conventional machine learning to physics-dominated manufacturing processes, where experimental data are scarce, expensive, and highly material-specific. The authors propose an integrated framework that combines physical knowledge with data-driven methods to systematically investigate, under limited-data conditions, the impact of data cleaning, feature selection, physics-informed fusion mechanisms, and evaluation strategies on model performance. They innovatively interpret statistically driven feature selection as an explicit modeling assumption and uncover significant instability in model evaluation under small-sample regimes. Experimental results demonstrate that a Gaussian process augmented with residual learning consistently outperforms other approaches, achieving well-calibrated predictive uncertainty—exhibiting an empirical 86% coverage for a nominal 90% prediction interval. However, while residual learning enhances the stability of Gaussian processes, it adversely affects tree-based models.

0 citationsRead paper

Directional Curvature from Armijo Backtracking: A Low-Cost Sharpness Probe and a Calibration-Free Learning-Rate Safeguard for Adam

Jul 04, 2026

In deep learning, the sharpness of a loss function—quantified by the largest eigenvalue of its Hessian—constrains the stability of gradient step sizes, yet conventional estimation methods are computationally prohibitive. This work reveals for the first time that the step size accepted by a single Armijo backtracking line search implicitly encodes directional curvature information, enabling an efficient approximation of sharpness. Building on this insight, we propose a low-overhead probe (≈1% computational cost) along the Adam update direction, combined with a fixed safety factor κ=2, to automatically set a calibration-free upper bound on the learning rate for Adam and AdamW. Experiments demonstrate strong negative correlations (−0.91 to −0.95) between logα and logλ₁ across CIFAR-10, Fashion-MNIST, and Imagenette, and show robust prevention of optimization divergence over initial learning rates spanning three orders of magnitude (10⁻³–3.0) without any additional hyperparameter tuning.

0 citationsRead paper

When Style Similarity Scores Fail: Diagnosing Raw CSD Cosine in Artist-Style Evaluation

May 09, 2026

This work addresses the common misuse of the original Cosine Similarity Distance (CSD) as an absolute measure of style fidelity, highlighting its lack of reliable calibration for artist-level style discrimination. The authors propose “discriminative gap”—a diagnostic method that requires neither prototypes nor thresholds—to evaluate CSD’s effectiveness in distinguishing intra- versus inter-style pairs. For the first time, they reveal shared limitations across mainstream vision backbones (CLIP, SigLIP, DINOv2) in style discrimination and introduce the CSD+ protocol, which integrates CSLS reweighting, positional embedding interpolation to 336 pixels, and a frozen backbone strategy. Experiments across 91 artists show that while the original CSD yields negative discriminative gaps for 15 artists, CSD+ reduces this to only 4 and improves unsupervised pairwise verification AUC from 0.883 to 0.905.

0 citationsRead paper

Multiple Additive Neural Networks for Structured and Unstructured Data

Apr 29, 2026

This work addresses the limitations of traditional gradient boosting methods—such as limited expressiveness, sensitivity to hyperparameters, and difficulty in continual learning—when handling both structured and unstructured data. The authors propose Multiplicative Additive Neural Networks (MANN), which replace decision trees with shallow neural networks as base learners to enable unified modeling of images, audio, and tabular data. Notably, MANN incorporates capsule networks for the first time into feature extraction for structured data. Embedded within an enhanced gradient boosting framework, MANN integrates a continual learning mechanism and regularization strategies, substantially reducing sensitivity to learning rate and iteration count. Experimental results demonstrate that MANN outperforms strong baselines such as XGBoost across multiple benchmark datasets, exhibiting superior generalization and multimodal adaptability.

0 citationsRead paper

Recent publications

Latest Papers

Multi-Class vs. Multi-Label BERT for CVE-to-CWE Mapping: How Taxonomy Structure Shapes the Errors

Jul 08, 2026

This study addresses the ongoing challenge of automatically mapping Common Vulnerabilities and Exposures (CVEs) to Common Weakness Enumerations (CWEs), particularly the lack of systematic evaluation in choosing between multi-class and multi-label modeling strategies. Framing the task as text classification, the authors compare these two approaches using three transformer-based encoders—BERT Base, SecureBERT, and CySecBERT—across CWE label spaces of varying granularity. They further introduce a hierarchical relaxed evaluation metric to more accurately assess model performance within the structured hierarchy of CWEs. Experimental results show that the multi-class approach generally achieves higher macro F1 scores, though this advantage diminishes as the label space narrows. CySecBERT performs best under the multi-label setting. Notably, the hierarchical relaxed metric elevates macro F1 from approximately 81% to 90%, indicating that most misclassifications stem from ambiguities in the CWE taxonomy rather than model limitations.

0 citationsRead paper

Physics-Informed Machine Learning Under Small-Data Constraints: Lessons from Abrasive Waterjet Milling

Jul 08, 2026

This study addresses the challenges of applying conventional machine learning to physics-dominated manufacturing processes, where experimental data are scarce, expensive, and highly material-specific. The authors propose an integrated framework that combines physical knowledge with data-driven methods to systematically investigate, under limited-data conditions, the impact of data cleaning, feature selection, physics-informed fusion mechanisms, and evaluation strategies on model performance. They innovatively interpret statistically driven feature selection as an explicit modeling assumption and uncover significant instability in model evaluation under small-sample regimes. Experimental results demonstrate that a Gaussian process augmented with residual learning consistently outperforms other approaches, achieving well-calibrated predictive uncertainty—exhibiting an empirical 86% coverage for a nominal 90% prediction interval. However, while residual learning enhances the stability of Gaussian processes, it adversely affects tree-based models.

0 citationsRead paper

Directional Curvature from Armijo Backtracking: A Low-Cost Sharpness Probe and a Calibration-Free Learning-Rate Safeguard for Adam

Jul 04, 2026

In deep learning, the sharpness of a loss function—quantified by the largest eigenvalue of its Hessian—constrains the stability of gradient step sizes, yet conventional estimation methods are computationally prohibitive. This work reveals for the first time that the step size accepted by a single Armijo backtracking line search implicitly encodes directional curvature information, enabling an efficient approximation of sharpness. Building on this insight, we propose a low-overhead probe (≈1% computational cost) along the Adam update direction, combined with a fixed safety factor κ=2, to automatically set a calibration-free upper bound on the learning rate for Adam and AdamW. Experiments demonstrate strong negative correlations (−0.91 to −0.95) between logα and logλ₁ across CIFAR-10, Fashion-MNIST, and Imagenette, and show robust prevention of optimization divergence over initial learning rates spanning three orders of magnitude (10⁻³–3.0) without any additional hyperparameter tuning.

0 citationsRead paper

When Style Similarity Scores Fail: Diagnosing Raw CSD Cosine in Artist-Style Evaluation

May 09, 2026

This work addresses the common misuse of the original Cosine Similarity Distance (CSD) as an absolute measure of style fidelity, highlighting its lack of reliable calibration for artist-level style discrimination. The authors propose “discriminative gap”—a diagnostic method that requires neither prototypes nor thresholds—to evaluate CSD’s effectiveness in distinguishing intra- versus inter-style pairs. For the first time, they reveal shared limitations across mainstream vision backbones (CLIP, SigLIP, DINOv2) in style discrimination and introduce the CSD+ protocol, which integrates CSLS reweighting, positional embedding interpolation to 336 pixels, and a frozen backbone strategy. Experiments across 91 artists show that while the original CSD yields negative discriminative gaps for 15 artists, CSD+ reduces this to only 4 and improves unsupervised pairwise verification AUC from 0.883 to 0.905.

0 citationsRead paper

Multiple Additive Neural Networks for Structured and Unstructured Data

Apr 29, 2026

This work addresses the limitations of traditional gradient boosting methods—such as limited expressiveness, sensitivity to hyperparameters, and difficulty in continual learning—when handling both structured and unstructured data. The authors propose Multiplicative Additive Neural Networks (MANN), which replace decision trees with shallow neural networks as base learners to enable unified modeling of images, audio, and tabular data. Notably, MANN incorporates capsule networks for the first time into feature extraction for structured data. Embedded within an enhanced gradient boosting framework, MANN integrates a continual learning mechanism and regularization strategies, substantially reducing sensitivity to learning rate and iteration count. Experimental results demonstrate that MANN outperforms strong baselines such as XGBoost across multiple benchmark datasets, exhibiting superior generalization and multimodal adaptability.

0 citationsRead paper