Selected work

Representative Papers

A Policy Algebra for Trust-Preserving Agentic AI Execution

Aug 17, 2026

This study addresses critical trust challenges in enterprise AI agents, including privilege escalation, data misuse, and audit deficiencies, by proposing a formal policy algebra framework. The approach defines reliable capability envelopes through compositional security constraints, integrating mechanisms for runtime obligation composition, budget tightening, and evidence accumulation to enable multi-agent trust propagation and cost-aware artifact materialization. Experimental evaluations demonstrate that the system intercepts 94.8% of violation events while achieving an 86.9% task completion rate and 98.6% audit integrity. These results indicate that the framework effectively eliminates critical security vulnerabilities, significantly enhancing both the reliability and auditability of agent execution while maintaining strict regulatory compliance.

0 citationsRead paper

An Investigation of the AUTOSAR Adaptive Platform from an Industry Perspective

Jul 06, 2026

This study addresses critical pain points in AUTOSAR Adaptive application development by revealing, for the first time from an industrial practice perspective, their root causes: inherent challenges arising from the interplay among the specification’s own architectural design and reuse objectives, vendor-specific implementation variations, and localized usage patterns. Employing a design science research methodology, the authors construct a minimal viable platform prototype and integrate configuration management analysis with runtime lifecycle modeling to systematically identify and attribute key issues. The primary contribution lies in demonstrating that design flaws at the specification level are the core catalysts of these challenges. The work further proposes optimizing the toolchain to reduce configuration complexity and training overhead, thereby offering empirical evidence and actionable pathways for improving the AUTOSAR Adaptive ecosystem.

0 citationsRead paper

Understanding Conversational Patterns in Multi-agent Programming: A Case Study on Fibonacci Game Development

May 22, 2026

This study addresses the challenges of multi-agent collaboration in software engineering—such as role misalignment, unstable convergence, and error propagation—that hinder reliable code generation. It presents the first systematic evaluation of dual-role agent collaboration (designer and programmer) along three dimensions: efficiency, consistency, and effectiveness. The authors construct 12 dialogue systems by pairing seven open-source large language models (Gemma 2/3, LLaMA 3.2/3.3, DeepSeek-R1, MiniCPM, and Qwen3) and conduct a multidimensional analysis using BLEU, ROUGE, and compilation success rates. Results show that DeepSeek-R1 self-pairing converges to the correct solution stably from the first round, while LLaMA 3.2 and Qwen3 self-pairings exhibit strong role alignment but deviate from correctness; all other pairings fail to converge effectively. This work provides a quantifiable framework and empirical insights for evaluating multi-agent programming collaboration.

0 citationsRead paper

BlendCLIP: Bridging Synthetic and Real Domains for Zero-Shot 3D Object Classification with Multimodal Pretraining

Oct 20, 2025

Zero-shot 3D object classification suffers from significant domain shift between synthetic LiDAR data and real-world sparse, noisy scans, severely limiting generalization in open-vocabulary settings. To address this, we propose a curriculum-based cross-domain fusion framework. First, we construct the first large-scale, real-scene point cloud–image–text triplet dataset. Second, we design a multimodal pretraining architecture that jointly leverages contrastive learning and curriculum learning to align point clouds, images, and text across domains. Our core innovation lies in synergistically integrating the semantic richness of synthetic data with the domain fidelity of real data, thereby mitigating domain gap and enhancing zero-shot transferability. On nuScenes, our method achieves 46.2% zero-shot accuracy—outperforming the prior state-of-the-art by 19.3 percentage points. It also attains new state-of-the-art results on outdoor benchmarks such as TruckScenes, demonstrating strong effectiveness and generalization under complex real-world conditions.

0 citationsRead paper

An AUTOSAR-Aligned Architectural Study of Vulnerabilities in Automotive SoC Software

Oct 09, 2025

In AUTOSAR-based automotive SoCs, real-time safety-critical software suffers from unclear vulnerability root causes and insufficient impact assessment. Method: We systematically analyze 180 publicly disclosed vulnerabilities to construct an SoC software architecture model aligned with the AUTOSAR layered reference architecture. We propose the first AUTOSAR-compliant vulnerability root cause taxonomy, identifying 16 root cause categories and 56 high-risk modules; further, we integrate CWE classification, layered abstraction modeling, and service-oriented analysis to empirically map vulnerabilities to architectural elements and quantify their cross-layer impacts and patching delays. Results: Our study reveals dominant vulnerability patterns, uncovers cross-layer distribution regularities and systematic patch delays, and provides actionable, architecture-level security enhancement guidelines—enabling vulnerability detection, prioritization, and precise localization. These findings support security-by-design optimization of automotive cyber-physical systems (CPS) platforms at the architectural level.

0 citationsRead paper

Recent publications

Latest Papers

A Policy Algebra for Trust-Preserving Agentic AI Execution

Aug 17, 2026

This study addresses critical trust challenges in enterprise AI agents, including privilege escalation, data misuse, and audit deficiencies, by proposing a formal policy algebra framework. The approach defines reliable capability envelopes through compositional security constraints, integrating mechanisms for runtime obligation composition, budget tightening, and evidence accumulation to enable multi-agent trust propagation and cost-aware artifact materialization. Experimental evaluations demonstrate that the system intercepts 94.8% of violation events while achieving an 86.9% task completion rate and 98.6% audit integrity. These results indicate that the framework effectively eliminates critical security vulnerabilities, significantly enhancing both the reliability and auditability of agent execution while maintaining strict regulatory compliance.

0 citationsRead paper

An Investigation of the AUTOSAR Adaptive Platform from an Industry Perspective

Jul 06, 2026

This study addresses critical pain points in AUTOSAR Adaptive application development by revealing, for the first time from an industrial practice perspective, their root causes: inherent challenges arising from the interplay among the specification’s own architectural design and reuse objectives, vendor-specific implementation variations, and localized usage patterns. Employing a design science research methodology, the authors construct a minimal viable platform prototype and integrate configuration management analysis with runtime lifecycle modeling to systematically identify and attribute key issues. The primary contribution lies in demonstrating that design flaws at the specification level are the core catalysts of these challenges. The work further proposes optimizing the toolchain to reduce configuration complexity and training overhead, thereby offering empirical evidence and actionable pathways for improving the AUTOSAR Adaptive ecosystem.

0 citationsRead paper

Understanding Conversational Patterns in Multi-agent Programming: A Case Study on Fibonacci Game Development

May 22, 2026

This study addresses the challenges of multi-agent collaboration in software engineering—such as role misalignment, unstable convergence, and error propagation—that hinder reliable code generation. It presents the first systematic evaluation of dual-role agent collaboration (designer and programmer) along three dimensions: efficiency, consistency, and effectiveness. The authors construct 12 dialogue systems by pairing seven open-source large language models (Gemma 2/3, LLaMA 3.2/3.3, DeepSeek-R1, MiniCPM, and Qwen3) and conduct a multidimensional analysis using BLEU, ROUGE, and compilation success rates. Results show that DeepSeek-R1 self-pairing converges to the correct solution stably from the first round, while LLaMA 3.2 and Qwen3 self-pairings exhibit strong role alignment but deviate from correctness; all other pairings fail to converge effectively. This work provides a quantifiable framework and empirical insights for evaluating multi-agent programming collaboration.

0 citationsRead paper

BlendCLIP: Bridging Synthetic and Real Domains for Zero-Shot 3D Object Classification with Multimodal Pretraining

Oct 20, 2025

Zero-shot 3D object classification suffers from significant domain shift between synthetic LiDAR data and real-world sparse, noisy scans, severely limiting generalization in open-vocabulary settings. To address this, we propose a curriculum-based cross-domain fusion framework. First, we construct the first large-scale, real-scene point cloud–image–text triplet dataset. Second, we design a multimodal pretraining architecture that jointly leverages contrastive learning and curriculum learning to align point clouds, images, and text across domains. Our core innovation lies in synergistically integrating the semantic richness of synthetic data with the domain fidelity of real data, thereby mitigating domain gap and enhancing zero-shot transferability. On nuScenes, our method achieves 46.2% zero-shot accuracy—outperforming the prior state-of-the-art by 19.3 percentage points. It also attains new state-of-the-art results on outdoor benchmarks such as TruckScenes, demonstrating strong effectiveness and generalization under complex real-world conditions.

0 citationsRead paper

An AUTOSAR-Aligned Architectural Study of Vulnerabilities in Automotive SoC Software

Oct 09, 2025

In AUTOSAR-based automotive SoCs, real-time safety-critical software suffers from unclear vulnerability root causes and insufficient impact assessment. Method: We systematically analyze 180 publicly disclosed vulnerabilities to construct an SoC software architecture model aligned with the AUTOSAR layered reference architecture. We propose the first AUTOSAR-compliant vulnerability root cause taxonomy, identifying 16 root cause categories and 56 high-risk modules; further, we integrate CWE classification, layered abstraction modeling, and service-oriented analysis to empirically map vulnerabilities to architectural elements and quantify their cross-layer impacts and patching delays. Results: Our study reveals dominant vulnerability patterns, uncovers cross-layer distribution regularities and systematic patch delays, and provides actionable, architecture-level security enhancement guidelines—enabling vulnerability detection, prioritization, and precise localization. These findings support security-by-design optimization of automotive cyber-physical systems (CPS) platforms at the architectural level.

0 citationsRead paper