REFINE: A Multi-Agent LLM Approach for Evidence-Guided Code Refactoring
本文提出REFINE，一种结合多代理和大语言模型的方法，用于生成Java代码重构候选方案，旨在减少代码质量问题同时避免引入新问题。
0 citationsRead paper
Selected work
Representative Papers
Resource-Constrained Affect Modelling via Variance Regularisation Pruning
This work addresses a critical limitation in existing model pruning methods for resource-constrained affective computing: their exclusive focus on sparsity while neglecting the impact of parameter removal on cross-user performance stability. To remedy this, the authors propose Variance-Regularized (VR) pruning, a novel framework that explicitly incorporates cross-user robustness into the pruning criterion. By jointly optimizing prediction accuracy and inter-subject performance variance, VR prioritizes retaining parameters that are robust to distributional shifts across users. The method employs a connection-sensitivity-based pruning strategy and is evaluated using the Concordance Correlation Coefficient (CCC) on the AGAIN dataset. Experimental results demonstrate that, even at an 80% pruning ratio without fine-tuning, the approach maintains competitive CCC performance, significantly enhancing the generalization capability and deployment feasibility of lightweight affective models in real-world interactive scenarios.
0 citationsRead paper
CyBOKClaw: Human-in-the-Loop CyBOK Mapping for Cybersecurity Curriculum
This study addresses the challenge of mapping cybersecurity course keywords to the Cybersecurity Body of Knowledge (CyBOK) due to ambiguous, overly broad terminology and incomplete alignment. To overcome this, the authors propose an interpretable human-in-the-loop retrieval framework that employs multi-level semantic strategies—including query normalization, manual term expansion, concept weight enhancement, enriched topic descriptions, and domain-sensitive ranking—to generate Top-k candidate CyBOK entries for expert review, thereby avoiding reliance on strict exact matching. The work introduces the ECA-5 evaluation metric, which assesses mapping utility based on expert judgment. Experimental results demonstrate that the proposed approach achieves a 98.00% ECA-5 accuracy on the validation set, significantly outperforming purely structural matching methods and effectively supporting experts in performing efficient and reliable knowledge alignment.
0 citationsRead paper
BYOT-CPS: A Hybrid Cyber-Physical Systems Testbed for IoT Security Assessment and Platform Evaluation
This study addresses the methodological gap in IoT security research between low-fidelity simulations and costly, hard-to-reproduce physical testbeds. To bridge this divide, the authors propose BYOT-CPS, a hybrid cyber-physical testbed that integrates real IoT devices—such as smart bulbs and cameras—with virtual networks emulated in GNS3. Designed to meet core requirements of fidelity, heterogeneity, scalability, reproducibility, and isolation, the platform implements a structured experimental environment comprising enterprise, service, attack, and monitoring zones. The system successfully demonstrates mixed physical-virtual networking, penetration testing, Mirai-style DDoS attack emulation, and fine-grained traffic monitoring. BYOT-CPS effectively narrows the gap between simulation and physical experimentation, offering a robust infrastructure for IoT security research, education, and third-party evaluation.
0 citationsRead paper
Gen-AI-tecture: using generative AI to support architectural students in design tasks
This study investigates how locally deployed, domain-specific generative AI tools can effectively enhance architectural students’ creativity, foster inclusive learning, and develop their AI literacy during design tasks. Innovatively embedding generative AI deeply into architectural education, the research employs a mixed-methods approach with focus group designs to empirically examine the synergistic mechanisms among constructivist, connectivist, and Universal Design for Learning theories within AI-augmented pedagogy. Findings demonstrate that this integrated approach significantly improves students’ creative fluency, increases engagement among learners from diverse backgrounds, and strengthens their confidence in AI-assisted design. The work thus offers both a theoretical foundation and a practical framework for the meaningful integration of AI in architectural education.
0 citationsRead paper
Recent publications
Latest Papers
REFINE: A Multi-Agent LLM Approach for Evidence-Guided Code Refactoring
本文提出REFINE，一种结合多代理和大语言模型的方法，用于生成Java代码重构候选方案，旨在减少代码质量问题同时避免引入新问题。
0 citationsRead paper
Resource-Constrained Affect Modelling via Variance Regularisation Pruning
This work addresses a critical limitation in existing model pruning methods for resource-constrained affective computing: their exclusive focus on sparsity while neglecting the impact of parameter removal on cross-user performance stability. To remedy this, the authors propose Variance-Regularized (VR) pruning, a novel framework that explicitly incorporates cross-user robustness into the pruning criterion. By jointly optimizing prediction accuracy and inter-subject performance variance, VR prioritizes retaining parameters that are robust to distributional shifts across users. The method employs a connection-sensitivity-based pruning strategy and is evaluated using the Concordance Correlation Coefficient (CCC) on the AGAIN dataset. Experimental results demonstrate that, even at an 80% pruning ratio without fine-tuning, the approach maintains competitive CCC performance, significantly enhancing the generalization capability and deployment feasibility of lightweight affective models in real-world interactive scenarios.
0 citationsRead paper
CyBOKClaw: Human-in-the-Loop CyBOK Mapping for Cybersecurity Curriculum
This study addresses the challenge of mapping cybersecurity course keywords to the Cybersecurity Body of Knowledge (CyBOK) due to ambiguous, overly broad terminology and incomplete alignment. To overcome this, the authors propose an interpretable human-in-the-loop retrieval framework that employs multi-level semantic strategies—including query normalization, manual term expansion, concept weight enhancement, enriched topic descriptions, and domain-sensitive ranking—to generate Top-k candidate CyBOK entries for expert review, thereby avoiding reliance on strict exact matching. The work introduces the ECA-5 evaluation metric, which assesses mapping utility based on expert judgment. Experimental results demonstrate that the proposed approach achieves a 98.00% ECA-5 accuracy on the validation set, significantly outperforming purely structural matching methods and effectively supporting experts in performing efficient and reliable knowledge alignment.
0 citationsRead paper
BYOT-CPS: A Hybrid Cyber-Physical Systems Testbed for IoT Security Assessment and Platform Evaluation
This study addresses the methodological gap in IoT security research between low-fidelity simulations and costly, hard-to-reproduce physical testbeds. To bridge this divide, the authors propose BYOT-CPS, a hybrid cyber-physical testbed that integrates real IoT devices—such as smart bulbs and cameras—with virtual networks emulated in GNS3. Designed to meet core requirements of fidelity, heterogeneity, scalability, reproducibility, and isolation, the platform implements a structured experimental environment comprising enterprise, service, attack, and monitoring zones. The system successfully demonstrates mixed physical-virtual networking, penetration testing, Mirai-style DDoS attack emulation, and fine-grained traffic monitoring. BYOT-CPS effectively narrows the gap between simulation and physical experimentation, offering a robust infrastructure for IoT security research, education, and third-party evaluation.
0 citationsRead paper
Gen-AI-tecture: using generative AI to support architectural students in design tasks
This study investigates how locally deployed, domain-specific generative AI tools can effectively enhance architectural students’ creativity, foster inclusive learning, and develop their AI literacy during design tasks. Innovatively embedding generative AI deeply into architectural education, the research employs a mixed-methods approach with focus group designs to empirically examine the synergistic mechanisms among constructivist, connectivist, and Universal Design for Learning theories within AI-augmented pedagogy. Findings demonstrate that this integrated approach significantly improves students’ creative fluency, increases engagement among learners from diverse backgrounds, and strengthens their confidence in AI-assisted design. The work thus offers both a theoretical foundation and a practical framework for the meaningful integration of AI in architectural education.
0 citationsRead paper