A Cubic-regularized Policy Newton Algorithm for Reinforcement Learning.
Mizhaan Prajit ManiyarAkash MondalPrashanth L. A.Shalabh BhatnagarPublished in: CoRR (2023)
Keyphrases
- learning algorithm
- reinforcement learning
- dynamic programming
- detection algorithm
- linear programming
- cost function
- optimization algorithm
- objective function
- computational cost
- matching algorithm
- primal dual
- significant improvement
- preprocessing
- computational complexity
- np hard
- lower bound
- stochastic approximation
- state space
- least squares
- worst case
- optimal solution
- image segmentation
- convergence analysis
- convergence rate
- policy search
- probabilistic model
- segmentation algorithm
- expectation maximization
- simulated annealing
- neural network
- machine learning