170 papers
The paper introduces SpecOps, a fully automated, multi-agent framework that effectively evaluates and detects bugs in real-world GUI-based AI agents by decomposing testing into specialized phases, outperforming existing baselines in accuracy, efficiency, and cost.
This paper introduces MALTA, a maintenance-aware scoring framework that significantly improves upon traditional Version Lag metrics by identifying high-risk, abandoned dependencies that would otherwise be overlooked, thereby addressing a critical blind spot in software supply chain security.
This paper introduces TOSSS, a CVE-based benchmark designed to evaluate the ability of Large Language Models to distinguish between secure and vulnerable code snippets in C/C++ and Java, revealing that current models achieve security scores ranging from 0.48 to 0.89.
This paper introduces DUCTILE, an agentic LLM orchestration framework that separates adaptive decision-making from deterministic tool execution to automate engineering analysis in product development, successfully handling input deviations in an aerospace case study while highlighting the emerging tension between task automation and the creation of exhausting supervisory roles.
This paper introduces FLA³, a governance-aware federated learning platform that integrates authentication, authorization, and accounting mechanisms to enable secure, privacy-preserving, and regulatory-compliant multi-center healthcare research, demonstrating its operational feasibility and clinical utility across international institutions while achieving predictive performance comparable to centralized training.
This paper introduces OAuthHub, a development framework that mitigates OAuth data overaccess by utilizing users' personal devices as intermediaries to enforce a centralized, just-in-time permission model, thereby significantly reducing both the code complexity and development time required for secure data sharing.
This paper introduces Agentic AIBOMs, a multi-agent framework that extends static Software Bills of Materials (SBOMs) with autonomous, policy-constrained reasoning to dynamically capture runtime behavior and environmental drift, thereby enhancing supply-chain security through reproducible, context-aware vulnerability assessment and minimal schema extensions to existing standards.
This paper presents and evaluates five prompt engineering strategies for reducing LLM hallucinations in industrial settings without modifying model weights, finding that an Enhanced Data Registry (M4) achieved perfect consistency in initial trials while a revised Decomposed Model-Agnostic Prompting (M2) showed the most significant improvement in subsequent verification.
This large-scale controlled study reveals that evaluation format (multiple-choice vs. open-ended) and specific model-scaffold interactions, rather than scaffold architecture alone, are the primary drivers of measured safety differences in language models, ultimately demonstrating that no universal safety ranking exists across different deployment configurations.
This paper proposes Nurture-First Development (NFD), a paradigm that shifts AI agent creation from static engineering to a continuous, conversational co-evolution process with domain experts, utilizing a Knowledge Crystallization Cycle to progressively transform tacit operational dialogue into structured, reusable expertise.
LabConstrictor is a GitHub-based automation tool that streamlines the deployment of Jupyter notebooks into one-click, installable desktop applications, thereby overcoming common barriers to software adoption and reproducibility in life sciences research.
This paper presents the first comprehensive evaluation of parameter-efficient fine-tuning (PEFT) for multitask code analysis, demonstrating that a single shared PEFT module can match or surpass full fine-tuning performance while significantly reducing computational and storage costs, provided that tasks are strategically grouped based on factors like complementarity and stability.
This paper addresses the inadequacy of traditional safety-case frameworks for modern AI systems by introducing comprehensive taxonomies of claims, arguments, and evidence, along with a reusable, structured template designed to construct credible and adaptive safety cases for generative and agentic AI.
UniCoR is a novel self-supervised framework that addresses the challenges of insufficient semantic understanding, inefficient modality fusion, and weak cross-language generalization in hybrid code retrieval by employing multi-perspective supervised contrastive learning and representation distribution consistency, thereby achieving state-of-the-art performance on both empirical and large-scale benchmarks.
This paper proposes a structured methodology that integrates the Robot Operating System (ROS) with Model-Based Systems Engineering (MBSE) through a specialized SysML metamodel called MeROS and an adapted V-model, aiming to enhance the semantic coherence, structural traceability, and reliable coordination of complex heterogeneous robotic systems.
This paper presents a systematic literature review of 80 studies on "systems of twinned systems," a paradigm that integrates the System of Systems and Digital Twin concepts to address modern engineering complexity, resulting in a new classification framework compatible with existing theories.
This paper presents the first study on JIT compiler performance bugs, combining an empirical analysis of 191 bug reports to identify common patterns with the development of "Jittery," a tool using layered differential performance testing that successfully discovered and helped fix multiple previously unknown performance issues in Oracle HotSpot and Graal compilers.
This paper introduces LoCoEval, the first benchmark designed to evaluate long-horizon conversational context management in repository-oriented development, revealing significant limitations in existing methods and proposing a unified memory approach that outperforms current baselines.
This paper presents a generalized feature model for Digital Twins, Digital Models, and Digital Shadows, derived from a systematic literature review and validated across emergency, vehicular, and manufacturing use cases, to facilitate informed design decisions, model-driven development, and verification.
This study demonstrates that large language models can effectively predict software story points without training data or with only a few examples, outperforming traditional supervised deep learning models, while also finding that comparative judgments, though not inherently easier to predict, can serve as effective few-shot examples to further enhance estimation accuracy.