Lyapunov-Driven Deep Reinforcement Learning for Edge Inference Empowered by Reconfigurable Intelligent Surfaces.
Kyriakos StylianopoulosMattia MerluzziPaolo Di LorenzoGeorge C. AlexandropoulosPublished in: ICASSP (2023)
Keyphrases
- reinforcement learning
- probabilistic inference
- adaptive control
- asymptotic stability
- closed loop
- dynamical systems
- edge information
- data driven
- surface reconstruction
- edge detection
- learning capabilities
- low cost
- bayesian networks
- function approximation
- three dimensional
- high curvature
- stability analysis
- learning algorithm
- sufficient conditions
- bayesian inference
- state space
- free form
- range data
- class of nonlinear systems
- robust stability
- learning process
- control theory
- intelligent systems
- model free
- optimal policy
- belief networks
- reinforcement learning algorithms
- action selection
- lyapunov function
- hardware implementation
- machine learning
- exponential stability
- edge detector