Constructs k-inductive neural barrier certificates for partially unknown nonlinear dynamics by combining neural networks, a data-driven fundamental lemma from one trajectory, and CEGIS-SMT verification.
arXiv preprint arXiv:2502.05510 , year=
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
representative citing papers
A single CBF-like constraint in quadratic optimization enables formation tracking for heterogeneous multi-agent systems using only relative neighbor information.
A neural-network barrier certificate over training trajectories certifies ℓ_p-bounded data-poisoning and evasion budgets, with PAC-style confidence, on MNIST, SVHN, and CIFAR-10.
citing papers explorer
-
k-Inductive Neural Barrier Certificates for Unknown Nonlinear Dynamics
Constructs k-inductive neural barrier certificates for partially unknown nonlinear dynamics by combining neural networks, a data-driven fundamental lemma from one trajectory, and CEGIS-SMT verification.
-
Control Barrier Function only Formation Tracking in Multi-Agent Systems
A single CBF-like constraint in quadratic optimization enables formation tracking for heterogeneous multi-agent systems using only relative neighbor information.
-
Robustness Certificates for Neural Networks Against Data Poisoning and Evasion Attacks
A neural-network barrier certificate over training trajectories certifies ℓ_p-bounded data-poisoning and evasion budgets, with PAC-style confidence, on MNIST, SVHN, and CIFAR-10.