Framework for verified quadratic characterizations of scalar relations via convex QPs and SOS certificates, applied to improve QC-based reachability analysis of feedforward neural networks with smooth activations and ReLU.
NNV: The neural network verification tool for deep neural networks and learning-enabled cyber- physical systems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Quadratic Characterizations for Reachability Analysis of Neural Networks
Framework for verified quadratic characterizations of scalar relations via convex QPs and SOS certificates, applied to improve QC-based reachability analysis of feedforward neural networks with smooth activations and ReLU.