ATLAS: AI-Assisted Threat-to-Assertion Learning for System-on-Chip Security Verification

ATLAS is a new AI-driven framework that uses large language models to automate System-on-Chip (SoC) security verification by converting threat models into formal verification properties. The system successfully detected 39 out of 48 security weaknesses in benchmark tests and generated correct security properties for 33 of those vulnerabilities.

• →ATLAS bridges standardized threat modeling with formal verification for System-on-Chip security using LLM technology.
• →The framework automatically transforms vulnerability knowledge from databases like CWE into testable security assertions.
• →Testing on HACK@DAC benchmarks showed 81% detection rate for security weaknesses with 85% accuracy in property generation.
• →The system advances toward secure-by-design paradigms by automating knowledge-driven security verification.
• →ATLAS generates JasperGold scripts for verification, enabling practical implementation in chip design workflows.