Lazily Adapted Constant Kinky Inference for Nonparametric Regression and Model-Reference Adaptive Control

Techniques known as Nonlinear Set Membership prediction, Lipschitz Interpolation or Kinky Inference are approaches to machine learning that utilise presupposed Lipschitz properties to compute inferences over unobserved function values. Provided a bound on the true best Lipschitz constant of the target function is known a priori they offer convergence guarantees as well as bounds around the predictions. Considering a more general setting that builds on Hoelder continuity relative to pseudo-metrics, we propose an online method for estimating the Hoelder constant online from function value observations that possibly are corrupted by bounded observational errors. Utilising this to compute adaptive parameters within a kinky inference rule gives rise to a nonparametric machine learning method, for which we establish strong universal approximation guarantees. That is, we show that our prediction rule can learn any continuous function in the limit of increasingly dense data to within a worst-case error bound that depends on the level of observational uncertainty. We apply our method in the context of nonparametric model-reference adaptive control (MRAC). Across a range of simulated aircraft roll-dynamics and performance metrics our approach outperforms recently proposed alternatives that were based on Gaussian processes and RBF-neural networks. For discrete-time systems, we provide guarantees on the tracking success of our learning-based controllers both for the batch and the online learning setting.

[1]  G. S. Watson,et al.  Smooth regression analysis , 1964 .

[2]  C. D. Perttunen,et al.  Lipschitzian optimization without the Lipschitz constant , 1993 .

[3]  David Muñoz de la Peña,et al.  Output feedback MPC based on smoothed projected kinky inference , 2019, IET Control Theory & Applications.

[4]  Gleb Beliakov Smoothing Lipschitz functions , 2007, Optim. Methods Softw..

[5]  E. Nadaraya On Estimating Regression , 1964 .

[6]  Robert L. Smith,et al.  Optimal estimation of univariate black-box Lipschitz functions with upper and lower error bounds , 2003, Comput. Oper. Res..

[7]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[8]  B. P. Zhang,et al.  Estimation of the Lipschitz constant of a function , 1996, J. Glob. Optim..

[9]  Robert M. Sanner,et al.  Gaussian Networks for Direct Adaptive Control , 1991, 1991 American Control Conference.

[10]  Ali Esmaili,et al.  Probability and Random Processes , 2005, Technometrics.

[11]  B. Pasik-Duncan,et al.  Adaptive Control , 1996, IEEE Control Systems.

[12]  A. Saad Simulation and Analysis of Wing Rock Physics for a Generic Fighter Model with Three Degrees-of-Freedom , 2000 .

[13]  Carl E. Rasmussen,et al.  Gaussian Processes for Data-Efficient Learning in Robotics and Control , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[14]  Jan Peters,et al.  Model learning for robot control: a survey , 2011, Cognitive Processing.

[15]  Jooyoung Park,et al.  Universal Approximation Using Radial-Basis-Function Networks , 1991, Neural Computation.

[16]  Carl E. Rasmussen,et al.  Gaussian process dynamic programming , 2009, Neurocomputing.

[17]  G. Grimmett,et al.  Probability and random processes , 2002 .

[18]  S. Shankar Sastry,et al.  Provably safe and robust learning-based model predictive control , 2011, Autom..

[19]  Boris Hasselblatt,et al.  A First Course in Dynamics: with a Panorama of Recent Developments , 2003 .

[20]  Leslie G. Valiant,et al.  A theory of the learnable , 1984, STOC '84.

[21]  B. Shubert A Sequential Method Seeking the Global Maximum of a Function , 1972 .

[22]  Adam Krzyzak,et al.  A Distribution-Free Theory of Nonparametric Regression , 2002, Springer series in statistics.

[23]  W. Cleveland Robust Locally Weighted Regression and Smoothing Scatterplots , 1979 .

[24]  Thomas G. Dietterich What is machine learning? , 2020, Archives of Disease in Childhood.

[25]  Jan-P. Calliess,et al.  Lipschitz optimisation for Lipschitz Interpolation , 2017, 2017 American Control Conference (ACC).

[26]  Miroslav Krstic,et al.  Control of Wing Rock Motion Using Adaptive Feedback Linearization , 1996 .

[27]  Alexandre B. Tsybakov,et al.  Introduction to Nonparametric Estimation , 2008, Springer series in statistics.

[28]  Carlo Novara,et al.  Set Membership identification of nonlinear systems , 2004, Autom..

[29]  J. How,et al.  Nonparametric Adaptive Control of Time-varying Systems using Gaussian Processes , 2012 .

[30]  T. Faniran Numerical Solution of Stochastic Differential Equations , 2015 .

[31]  Girish Chowdhary,et al.  A Bayesian nonparametric approach to adaptive control using Gaussian Processes , 2013, 52nd IEEE Conference on Decision and Control.

[32]  Jonathan P. How,et al.  Bayesian Nonparametric Adaptive Control Using Gaussian Processes , 2015, IEEE Transactions on Neural Networks and Learning Systems.

[33]  Carl E. Rasmussen,et al.  PILCO: A Model-Based and Data-Efficient Approach to Policy Search , 2011, ICML.

[34]  G. Beliakov Interpolation of Lipschitz functions , 2006 .

[35]  Erik D. Demaine,et al.  Optimally Adaptive Integration of Univariate Lipschitz Functions , 2006, Algorithmica.

[36]  G. G. Stokes "J." , 1890, The New Yale Book of Quotations.

[37]  David S. Broomhead,et al.  Multivariable Functional Interpolation and Adaptive Networks , 1988, Complex Syst..

[38]  Massimo Canale,et al.  Nonlinear model predictive control from data: a set membership approach , 2014 .

[39]  Karl Johan Åström,et al.  Adaptive Control , 1989, Embedded Digital Control with Microcontrollers.

[40]  Duane A. Cooper Learning lipschitz functions , 1995, Int. J. Comput. Math..

[41]  Jan Peters,et al.  A Survey on Policy Search for Robotics , 2013, Found. Trends Robotics.

[42]  Frank L. Lewis,et al.  High-Level Feedback Control with Neural Networks , 1998, World Scientific Series in Robotics and Intelligent Systems.

[43]  Francisco Curbera,et al.  Optimal Integration of Lipschitz Functions with a Gaussian Weight , 1998, J. Complex..

[44]  C. Gardiner Stochastic Methods: A Handbook for the Natural and Social Sciences , 2009 .

[45]  A. G. Sukharev Optimal method of constructing best uniform approximations for functions of a certain class , 1978 .

[46]  Steffen Dereich,et al.  Infinite-Dimensional Quadrature and Quantization , 2006 .

[47]  Radu Miculescu Approximation of Continuous Functions by Lipschitz Functions , 2000 .

[48]  Stephen J. Roberts,et al.  Conservative decision-making and interference in uncertain dynamical systems , 2014 .