Scalable Gradient Ascent for Controllers in Constrained POMDPs.
Kyle Hollins WrayKenneth CzuprynskiPublished in: ICRA (2022)
Keyphrases
- gradient ascent
- partially observable markov decision processes
- policy gradient
- reinforcement learning
- expectation maximization
- cross entropy
- exponential family
- control system
- conjugate gradient
- dynamic programming
- state space
- decision problems
- optimal control
- finite state
- planning problems
- function approximation
- control strategy
- markov decision processes
- em algorithm
- probability distribution
- multi agent systems