Geometry and convergence of natural policy gradient methods
From MaRDI portal
Publication:6138809
DOI10.1007/s41884-023-00106-zarXiv2211.02105OpenAlexW4379185386MaRDI QIDQ6138809
No author found.
Publication date: 16 January 2024
Published in: Information Geometry (Search for Journal in Brave)
Full work available at URL: https://arxiv.org/abs/2211.02105
Markov decision processHessian geometrystochastic policynatural policy gradientstate-action frequency
Cites Work
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- On the Fisher metric of conditional probability polytopes
- Information geometry and its applications
- Wasserstein Riemannian geometry of Gaussian densities
- Natural gradient via optimal transport
- Simple statistical gradient-following algorithms for connectionist reinforcement learning
- Hessian informed mirror descent
- Finite state Markovian decision processes
- Axiomatic Geometry of Conditional Models
- An Extended Cencov Characterization of the Information Metric
- Inexact Newton Methods
- Survey of linear programming for standard and nonstandard Markovian control problems. Part I: Theory
- OnActor-Critic Algorithms
- Simulation-based optimization of Markov reward processes
- Hessian Riemannian Gradient Flows in Convex Programming
- Fast Global Convergence of Natural Policy Gradient Methods with Entropy Regularization
- Global Convergence of Policy Gradient Methods to (Almost) Locally Optimal Policies
- Derivatives of Logarithmic Stationary Distributions for Policy Gradient Reinforcement Learning
- Information Geometry
- Approximate Newton Policy Gradient Algorithms
- Invariance properties of the natural gradient in overparametrised systems
- Efficient Natural Gradient Descent Methods for Large-Scale PDE-Based Optimization Problems
- Policy Mirror Descent for Regularized Reinforcement Learning: A Generalized Framework with Linear Convergence
This page was built for publication: Geometry and convergence of natural policy gradient methods
