Paper #CONF-105 Auto-Save Active (Saved ✓)
Inline Manuscript PreviewPDF Viewer (v1.0)
Automated Verification of Neural Network Constraints in Real-Time Systems
[Double-Blind Submission — Author names anonymized]
Abstract
Neural network verification guarantees safety properties in autonomous real-time applications. In this paper, we propose a scalable automated verification framework capable of evaluating constraint bounds across non-linear activation layers...
1. Introduction
Real-time safety critical systems rely increasingly on machine learning components. Verifying bounds is paramount...