A Meta-Learning Control Algorithm with Provable Finite-Time Guarantees
arXiv:2008.13265
Abstract
In this work we provide provable regret guarantees for an online meta-learning control algorithm in an iterative control setting, where in each iteration the system to be controlled is a linear deterministic system that is different and unknown, the cost for the controller in an iteration is a general additive cost function and the control input is required to be constrained, which if violated incurs an additional cost. We prove (i) that the algorithm achieves a regret for the controller cost and constraint violation that are for an episode of duration with respect to the best policy that satisfies the control input control constraints and (ii) that the average of the regret for the controller cost and constraint violation with respect to the same policy vary as with the number of iterations , showing that the worst regret for the learning within an iteration continuously improves with experience of more iterations.
arXiv commentThe problem formulation considered has a trivial solution and does not require the elaborate approach proposed in this paper. Moreover the assumptions made are very restrictive. We direct the reader to our next version titled: 'Meta-Learning Guarantees for Online Receding Horizon Learning Control'
References in corpus (12)
- A Simple Neural Attentive Meta-Learner
- RL: Fast Reinforcement Learning via Slow Reinforcement Learning
- Learning to reinforcement learn
- Global Convergence of Policy Gradient Methods for the Linear Quadratic Regulator
- Meta-Learning with Implicit Gradients
- Learning to Optimize
- Online Meta-Learning
- Online convex optimization for cumulative constraints
- Online Control with Adversarial Disturbances
- Logarithmic Regret for Online Control
- Provable Guarantees for Gradient-Based Meta-Learning
- A Stochastic View of Optimal Regret through Minimax Duality