analiza aiqbench
GuardEn introduces a novel framework that translates complex visual safety policies into executable code, enabling interpretable reasoning grounded in scene graphs rather than relying solely on black-box training. By decomposing safety rules into atomic propositions, this approach significantly improves adaptability and explainability for evolving risk patterns in vision-language models. The method demonstrates substantial gains on SafetyVisionBench, offering a programmable path toward more robust AI safety systems.
Visual Compliance via Executable Safety Rule Entailment
arXiv:2609.18328v1 Announce Type: new Abstract: Recent advances in LLMs and VLMs have enabled safety systems to reason beyond simple risk patterns toward more contextual and semantic safety concerns. However, as risk patterns continue to evolve and safety rules become more complex, existing training-based end-to-end safeguards face persistent challenges in adaptability and explainable reasoning over complex safety rules. To address these challenges, we propose GuardEn (Guarding by Safety Rule Entailment), an executable safeguard framework that decomposes safety policies into atomic propositions through Safety-Rule Compilation, modeling their composition as executable code. At test time, Scene-Grounded Execution instantiates these atomic propositions with contextual visual information derived from scene graphs, enabling rule-grounded and interpretable safety reasoning. Experiments on SafetyVisionBench demonstrate the effectiveness of programmable safeguard for complex visual safety assessment, achieving an average improvement of 9.8 F1 points over the strongest baseline.
Abstract:Recent advances in LLMs and VLMs have enabled safety systems to reason beyond simple risk patterns toward more contextual and semantic safety concerns. However, as risk patterns continue to evolve and safety rules become more complex, existing training-based end-to-end safeguards face persistent challenges in adaptability and explainable reasoning over complex safety rules. To address these challenges, we propose GuardEn (Guarding by Safety Rule Entailment), an executable safeguard framework that decomposes safety policies into atomic propositions through Safety-Rule Compilation, modeling their composition as executable code. At test time, Scene-Grounded Execution instantiates these atomic propositions with contextual visual information derived from scene graphs, enabling rule-grounded and interpretable safety reasoning. Experiments on SafetyVisionBench demonstrate the effectiveness of programmable safeguard for complex visual safety assessment, achieving an average improvement of 9.8 F1 points over the strongest baseline.
Accepted to EMNLP 2026. 36 pages, 20 figures, 24 tables
Artificial Intelligence (cs.AI)
arXiv:2609.18328 [cs.AI]
(or arXiv:2609.18328v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2609.18328 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 237 cuvinte
Subiecte conexe
Vertical Advantage: Transforming Industries with Lakebase and Agentic AI
Databricks · 27 aug. 2026
Links #5: 2026/07
LessWrong · 31 iul. 2026
Hands-free first notice of loss: Using Strands Agents and Amazon Bedrock AgentCore Browser Tool for intelligent claims intake
AWS · 9 iun. 2026
A shared agentic platform for Wood Mackenzie, on Amazon Bedrock AgentCore
AWS · 17 sept. 2026