A constrained symbolic regression method on expression trees discovers Lyapunov functions for autonomous dynamical systems without assuming their functional form.
EDMD-basedrobustobserversynthesisfornonlinearsystems
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
years
2026 4verdicts
UNVERDICTED 4representative citing papers
Neural networks learn the dynamics and mapping of an extended KKL observer for nonautonomous nonlinear systems from data, enabling state observation with a proven error bound on new inputs.
Nonlinear discrete-time systems are shown to admit exact bilinear representations via separate RKHS lifts of state and input, with stabilization posed as optimization over conditional probability measures.
The paper develops three convex learning settings for hybrid models that enforce interpretability via reference regularization, subspace restrictions, and nonlinear manifold restrictions, re-parameterized through lifted operator features as kernel mixtures of interpretable components.
citing papers explorer
-
A constrained symbolic regression approach for Lyapunov function discovery
A constrained symbolic regression method on expression trees discovers Lyapunov functions for autonomous dynamical systems without assuming their functional form.
-
Neural Luenberger state observer for nonautonomous nonlinear systems
Neural networks learn the dynamics and mapping of an extended KKL observer for nonautonomous nonlinear systems from data, enabling state observation with a proven error bound on new inputs.
-
Koopman Modeling and Stabilization of Discrete-Time Nonlinear Control Systems: Bilinearity on a Reproducing Kernel Hilbert Space
Nonlinear discrete-time systems are shown to admit exact bilinear representations via separate RKHS lifts of state and input, with stabilization posed as optimization over conditional probability measures.
-
Convex Hybrid Modeling: An Operator-Based Approach
The paper develops three convex learning settings for hybrid models that enforce interpretability via reference regularization, subspace restrictions, and nonlinear manifold restrictions, re-parameterized through lifted operator features as kernel mixtures of interpretable components.