Implicit Regularization and Momentum Algorithms in Nonlinearly Parameterized Adaptive Control and Prediction
From MaRDI portal
Abstract: Stable concurrent learning and control of dynamical systems is the subject of adaptive control. Despite being an established field with many practical applications and a rich theory, much of the development in adaptive control for nonlinear systems revolves around a few key algorithms. By exploiting strong connections between classical adaptive nonlinear control techniques and recent progress in optimization and machine learning, we show that there exists considerable untapped potential in algorithm development for both adaptive nonlinear control and adaptive dynamics prediction. We first introduce first-order adaptation laws inspired by natural gradient descent and mirror descent. We prove that when there are multiple dynamics consistent with the data, these non-Euclidean adaptation laws implicitly regularize the learned model. Local geometry imposed during learning thus may be used to select parameter vectors - out of the many that will achieve perfect tracking or prediction - for desired properties such as sparsity. We apply this result to regularized dynamics predictor and observer design, and as concrete examples consider Hamiltonian systems, Lagrangian systems, and recurrent neural networks. We subsequently develop a variational formalism based on the Bregman Lagrangian to define adaptation laws with momentum applicable to linearly parameterized systems and to nonlinearly parameterized systems satisfying monotonicity or convexity requirements. We show that the Euler Lagrange equations for the Bregman Lagrangian lead to natural gradient and mirror descent-like adaptation laws with momentum, and we recover their first-order analogues in the infinite friction limit. We illustrate our analyses with simulations demonstrating our theoretical results.
Recommendations
- Implicit predictive adaptive control with stochastic gradient identifiers
- Adaptive control with nonconvex parameterization
- scientific article; zbMATH DE number 932447
- Adaptive control vector parameterization for nonlinear model-predictive control
- scientific article; zbMATH DE number 970345
- Adaptive iterative learning control for <scp>2D</scp> nonlinear systems with nonrepetitive uncertainties
- Immersion and Invariance Adaptive Control of Nonlinearly Parameterized Nonlinear Systems
- Adaptive control of nonlinear systems via approximate linearization
- Adaptive iterative learning control for a class of non-parametric uncertain systems
- Approximate Nonlinear Regulation via Identification-Based Adaptive Internal Models
Cites work
- A Continuous-Time Analysis of Distributed Stochastic Gradient
- A differential equation for modeling Nesterov's accelerated gradient method: theory and insights
- A variational perspective on accelerated methods in optimization
- Acceleration of Stochastic Approximation by Averaging
- Adaptation algorithms in finite form for nonlinear dynamic objects
- Adaptation and Parameter Estimation in Systems With Unstable Target Dynamics and Nonlinear Parametrization
- Adaptation in the presence of a general nonlinear parameterization: an error model approach
- Adaptive control of continuous time systems with convex/concave parametrization
- Adaptive control of nonlinearly parameterized systems with a triangular structure
- Adaptive sliding controller synthesis for non-linear systems
- Benign overfitting in linear regression
- Conformal symplectic and relativistic optimization
- Data-driven discovery of coordinates and governing equations
- Dynamic scaling and observer design with application to adaptive control
- Generalized momentum-based methods: a Hamiltonian perspective
- scientific article; zbMATH DE number 5917373 (Why is no real title available?)
- scientific article; zbMATH DE number 3850830 (Why is no real title available?)
- scientific article; zbMATH DE number 3979891 (Why is no real title available?)
- scientific article; zbMATH DE number 3718924 (Why is no real title available?)
- scientific article; zbMATH DE number 3790208 (Why is no real title available?)
- scientific article; zbMATH DE number 53595 (Why is no real title available?)
- scientific article; zbMATH DE number 124710 (Why is no real title available?)
- scientific article; zbMATH DE number 1377399 (Why is no real title available?)
- scientific article; zbMATH DE number 845714 (Why is no real title available?)
- scientific article; zbMATH DE number 3313108 (Why is no real title available?)
- Immersion and Invariance Adaptive Control of Nonlinearly Parameterized Nonlinear Systems
- Immersion and invariance: a new tool for stabilization and adaptive control of nonlinear systems
- Mirror descent and nonlinear projected subgradient methods for convex optimization.
- Modular stability tools for distributed computation and control
- Observability of complex systems
- On contraction analysis for non-linear systems
- Parameter estimation of nonlinearly parameterized regressions without overparameterization: application to adaptive control
- Reconciling modern machine-learning practice and the classical bias-variance trade-off
- Some methods of speeding up the convergence of iteration methods
- Speed-gradient scheme and its application in adaptive control problems
- The Feynman lectures on physics 4. The structure of matter. Translated from the English by Henner Wessel and Karen Lippert
- The implicit bias of gradient descent on separable data
- The robustness of the \(p\)-norm algorithms
- Velocity gradient algorithms in control and adaptation problems
Cited in
(5)- Further results on implicit models with application to LQ adaptive optimization
- Speed gradient method and its applications
- Finite-memory convergent algorithm for estimating the regression function and its application in adaptive control
- Accelerated-gradient-approach-based adaptive deep neural optimized control for strict-feedback systems with dynamic-memory event-triggered mechanism
- Nonparametric adaptive payload tracking for an offshore crane
This page was built for publication: Implicit Regularization and Momentum Algorithms in Nonlinearly Parameterized Adaptive Control and Prediction
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5004318)
