Melanie Zeilinger: "Learning-based Model Predictive Control - Towards Safe Learning in Control"
Intersections between Control, Learning and Optimization 2020
"Learning-based Model Predictive Control - Towards Safe Learning in Control"
Melanie Zeilinger - ETH Zurich & University of Freiburg
Abstract: The question of safety when integrating learning techniques in control systems has been recognized as a central challenge for the widespread success of these promising techniques. While different notions of safety exist, I will focus on the satisfaction of critical safety constraints (in probability) in this talk, a common and intuitive form of specifying safety in many applications. Optimization-based control has been established as the main technique for systematically addressing constraint satisfaction in the control of complex systems. However, it suffers from the need of a mathematical problem representation, i.e. a model, constraints and objective. Reinforcement learning, in contrast, has demonstrated its success for complex problems where a mathematical problem representation is not available by directly interacting with the system, however, at the cost of safety guarantees.
In this talk, I will discuss techniques that aim at bridging these two paradigms. We will investigate three variants how learning can be combined with optimization-based concepts to generate high-performance controllers that are simple and time-efficient to design while offering a notion of constraint satisfaction, and thereby of safety. We will begin with techniques for inferring a model of the dynamics, objective or constraints from data for the integration in optimization-based control, and then discuss a safety filter as a modular approach for augmenting reinforcement learning with constraint satisfaction properties. I will show examples of using these techniques in robotics applications.
Institute for Pure and Applied Mathematics, UCLA
February 26, 2020
For more information: http://www.ipam.ucla.edu/lco2020
Видео Melanie Zeilinger: "Learning-based Model Predictive Control - Towards Safe Learning in Control" канала Institute for Pure & Applied Mathematics (IPAM)
"Learning-based Model Predictive Control - Towards Safe Learning in Control"
Melanie Zeilinger - ETH Zurich & University of Freiburg
Abstract: The question of safety when integrating learning techniques in control systems has been recognized as a central challenge for the widespread success of these promising techniques. While different notions of safety exist, I will focus on the satisfaction of critical safety constraints (in probability) in this talk, a common and intuitive form of specifying safety in many applications. Optimization-based control has been established as the main technique for systematically addressing constraint satisfaction in the control of complex systems. However, it suffers from the need of a mathematical problem representation, i.e. a model, constraints and objective. Reinforcement learning, in contrast, has demonstrated its success for complex problems where a mathematical problem representation is not available by directly interacting with the system, however, at the cost of safety guarantees.
In this talk, I will discuss techniques that aim at bridging these two paradigms. We will investigate three variants how learning can be combined with optimization-based concepts to generate high-performance controllers that are simple and time-efficient to design while offering a notion of constraint satisfaction, and thereby of safety. We will begin with techniques for inferring a model of the dynamics, objective or constraints from data for the integration in optimization-based control, and then discuss a safety filter as a modular approach for augmenting reinforcement learning with constraint satisfaction properties. I will show examples of using these techniques in robotics applications.
Institute for Pure and Applied Mathematics, UCLA
February 26, 2020
For more information: http://www.ipam.ucla.edu/lco2020
Видео Melanie Zeilinger: "Learning-based Model Predictive Control - Towards Safe Learning in Control" канала Institute for Pure & Applied Mathematics (IPAM)
Показать
Комментарии отсутствуют
Информация о видео
10 апреля 2020 г. 3:59:28
00:51:10
Другие видео канала
Richard Murray: "Can We Really Use Machine Learning in Safety Critical Systems?"Understanding Model Predictive Control, Part 1: Why Use MPC?Francesco Borrelli: "Sample-Based Learning Model Predictive Control"On Using Formal Methods For Safe and Robust Robot Autonomy (Karen Leung, PhD Defense)The secrets of learning a new language | Lýdia MachováDana Pe'er: "Having fun with single-cell RNA-seq: imputation and manifolds"Yann LeCun: "Energy-Based Self-Supervised Learning"Model Predictive ControlZico Kolter: "Integrating optimization, constraints, and control within deep learning models"Manfred Morari (University of Pennsylvania): "A Practitioner's Perspective"Model Predictive Control System | Neural Network | MATLAB Helper"Reinforcement Learning for Recommender Systems: A Case Study on Youtube," by Minmin ChenSteve Brunton: "Introduction to Fluid Mechanics"Introduction to Model Predictive Control ToolboxA Gentle Introduction to Offline Reinforcement LearningGerardo Bledt (MIT): Generalizing and improving regularized predictive control for legged robotsStéphane Mallat: "Scattering Invariant Deep Networks for Classification, Pt. 1"Reinforcement Learning: Machine Learning Meets Control TheoryBenjamin Recht: Optimization Perspectives on Learning to Control (ICML 2018 tutorial)