Online Learning of Nonlinear Parametric Models under Non-smooth Regularization using EKF and ADMM
By: Lapo Frascati, Alberto Bemporad
Potential Business Impact:
Teaches computers to learn from new data fast.
This paper proposes a novel combination of extended Kalman filtering (EKF) with the alternating direction method of multipliers (ADMM) for learning parametric nonlinear models online under non-smooth regularization terms, including l1 and l0 penalties and bound constraints on model parameters. For the case of linear time-varying models and non-smoothconvex regularization terms, we provide a sublinear regret bound that ensures the proper behavior of the online learning strategy. The approach is computationally efficient for a wide range of regularization terms, which makes it appealing for its use in embedded control applications for online model adaptation. We show the performance of the proposed method in three simulation examples, highlighting its effectiveness compared to other batch and online algorithms.
Similar Papers
Annealed Ensemble Kalman Inversion for Constrained Nonlinear Model Predictive Control: An ADMM Approach
Optimization and Control
Helps robots learn to move safely and efficiently.
Smoothing ADMM for Non-convex and Non-smooth Hierarchical Federated Learning
Machine Learning (CS)
Trains AI smarter and faster with different data.
Model-free Online Learning for the Kalman Filter: Forgetting Factor and Logarithmic Regret
Machine Learning (CS)
Makes predictions better for changing systems.