Selected work

Representative Papers

Tangentially Aligned Integrated Gradients for User-Friendly Explanations

Mar 11, 2025

To address the subjectivity and instability of baseline selection in Integrated Gradients (IG), which often leads to explanations misaligned with the underlying data manifold, this paper proposes an automatic baseline optimization method guided by **maximizing tangential alignment**. We formally define the alignment degree of explanation vectors within the Riemannian tangent space of the data manifold, derive the theoretical conditions under which IG vectors lie in this tangent space, and design a differentiable approximation enabling end-to-end optimization. Unlike conventional approaches relying on heuristic baselines (e.g., zero vector or dataset mean), our method requires no manual baseline specification and seamlessly integrates into the standard IG framework. Experiments on ImageNet and CIFAR-10 demonstrate substantial improvements in explanation consistency, stability, and human interpretability over zero-baseline IG, mean-baseline IG, and Grad-CAM, empirically validating the effectiveness of manifold-aware attribution.

1 citationsRead paper

Decoding Decision Correctness from EEG Under High Cognitive Workload in Virtual Reality: Implications for Collaborative Brain-Computer Interface Teams

Sep 02, 2026

研究通过EEG在高认知负荷下预测决策正确性，提出基于EEG的信号可在决策前提高团队决策准确性，尤其在高工作负荷条件下。

0 citationsRead paper

InSPECtor: Improving SLEIGH Processor Specification Veracity via Proxy

Aug 13, 2026

This work addresses the long-standing lack of systematic validation for processor specifications, which can lead to distorted program behavior and security vulnerabilities. It presents the first automated differential testing framework tailored for open-source SLEIGH specifications, automatically generating decodable instructions and initial execution states by parsing specification structures, and systematically validating them against multiple hardware reference implementations across architectures. Applied to x86-64 and AArch64, the approach uncovered 38,920 semantic discrepancies, identified 125 unique defects—many of which were subsequently fixed—and significantly improved specification fidelity. Furthermore, it exposed inconsistencies across vendor implementations and led to eight concrete recommendations, establishing a new paradigm for ensuring the reliability of instruction set architecture specifications.

0 citationsRead paper

Improving Interoperability among Defence and National Security Ontologies: Analysis and Evaluation Tasks

Aug 06, 2026

This study addresses the challenge of semantic interoperability in the defense and national security domain, where numerous highly heterogeneous and specialized ontologies impede effective integration. To bridge this gap, the authors systematically analyze over 60 publicly available ontologies and establish the first Ontology Alignment Evaluation Initiative (OAEI) benchmark track dedicated to this domain, comprising eight alignment tasks. Leveraging multiple state-of-the-art ontology matching systems, they generate automatic alignments, aggregate them into a consensus mapping, and refine the results through expert manual validation to produce a high-quality silver-standard dataset. This work fills a critical void in standardized evaluation for ontology alignment in defense and security contexts, significantly advancing semantic interoperability within the field.

0 citationsRead paper

Semantic Robustness Certification for Vision-Language Models

Jun 17, 2026

Existing vision-language models lack verifiable robustness guarantees under distribution shifts induced by semantic variations such as shape, size, or style. This work proposes the first framework capable of certifying robustness against semantic-level perturbations without requiring additional data. Leveraging the open-vocabulary capacity of vision-language models, the method employs textual prompts as semantic proxies to construct controllable transformations and derives closed-form characterizations of decision boundaries to quantitatively certify invariant prediction regions under semantic perturbations. By moving beyond the conventional limitations of geometric or pixel-level perturbations, the approach demonstrates effective and practical robustness certification across diverse semantic transformations on both synthetic and real-world datasets.

0 citationsRead paper

Recent publications

Latest Papers

Decoding Decision Correctness from EEG Under High Cognitive Workload in Virtual Reality: Implications for Collaborative Brain-Computer Interface Teams

Sep 02, 2026

研究通过EEG在高认知负荷下预测决策正确性，提出基于EEG的信号可在决策前提高团队决策准确性，尤其在高工作负荷条件下。

0 citationsRead paper

InSPECtor: Improving SLEIGH Processor Specification Veracity via Proxy

Aug 13, 2026

This work addresses the long-standing lack of systematic validation for processor specifications, which can lead to distorted program behavior and security vulnerabilities. It presents the first automated differential testing framework tailored for open-source SLEIGH specifications, automatically generating decodable instructions and initial execution states by parsing specification structures, and systematically validating them against multiple hardware reference implementations across architectures. Applied to x86-64 and AArch64, the approach uncovered 38,920 semantic discrepancies, identified 125 unique defects—many of which were subsequently fixed—and significantly improved specification fidelity. Furthermore, it exposed inconsistencies across vendor implementations and led to eight concrete recommendations, establishing a new paradigm for ensuring the reliability of instruction set architecture specifications.

0 citationsRead paper

Improving Interoperability among Defence and National Security Ontologies: Analysis and Evaluation Tasks

Aug 06, 2026

This study addresses the challenge of semantic interoperability in the defense and national security domain, where numerous highly heterogeneous and specialized ontologies impede effective integration. To bridge this gap, the authors systematically analyze over 60 publicly available ontologies and establish the first Ontology Alignment Evaluation Initiative (OAEI) benchmark track dedicated to this domain, comprising eight alignment tasks. Leveraging multiple state-of-the-art ontology matching systems, they generate automatic alignments, aggregate them into a consensus mapping, and refine the results through expert manual validation to produce a high-quality silver-standard dataset. This work fills a critical void in standardized evaluation for ontology alignment in defense and security contexts, significantly advancing semantic interoperability within the field.

0 citationsRead paper

Semantic Robustness Certification for Vision-Language Models

Jun 17, 2026

Existing vision-language models lack verifiable robustness guarantees under distribution shifts induced by semantic variations such as shape, size, or style. This work proposes the first framework capable of certifying robustness against semantic-level perturbations without requiring additional data. Leveraging the open-vocabulary capacity of vision-language models, the method employs textual prompts as semantic proxies to construct controllable transformations and derives closed-form characterizations of decision boundaries to quantitatively certify invariant prediction regions under semantic perturbations. By moving beyond the conventional limitations of geometric or pixel-level perturbations, the approach demonstrates effective and practical robustness certification across diverse semantic transformations on both synthetic and real-world datasets.

0 citationsRead paper

TriAlign: Towards Universal Truth Consistency in Personalized LLM Alignment

Jun 01, 2026

This work addresses the tension between personalization and fairness in large language models, where adapting to individual user preferences may compromise consistency and equity across social groups on objective factual tasks. To mitigate this issue, the authors propose Truth-Invariant Alignment (TIA), a novel alignment objective that preserves universal factual consistency while maintaining personalization capabilities. They introduce TriAlign, the first offline multi-agent reinforcement learning framework designed for TIA, which models distinct social groups as interacting agents and incorporates a fairness-aware optimization objective alongside an explicit inconsistency penalty. Experimental results demonstrate that TriAlign significantly reduces inter-group disparities in factual responses while simultaneously improving performance on objective tasks and retaining high-quality personalization, outperforming strong existing baselines.

0 citationsRead paper