SOURCE-LINKED INTELLIGENCE

Compute-Bounded Security Assurance - Coverage, Verification, and Response under Resource Constraints

arXiv · AI, language, vision and robotics · article · Sep 7, 2026 · UTC

Additional inference compute can increase the number of correctly resolved security-assurance tasks, but repeated success, unique coverage, accepted evidence, and operational protection are different quantities. We develop a resource-constrained framework that separates them. For repeated conditionally independent attempts with latent success probability $Θ$, coverage is $C_n = 1 - E[(1-Θ)^n]$, and its limiting value is $1 - P(Θ= 0)$. Positive pairwise outcome correlation does not by itself imply a ceiling below one: we construct two models with the same mean success and pairwise correlation b

Read original source ↗ Open in workspace

recordType

paper

region

Global