Differentiable Trust Region Layers for Deep Reinforcement Learning.
Fabian OttoPhilipp BeckerNgo Anh VienHanna Carolin Maria ZiescheGerhard NeumannPublished in: ICLR (2021)
Keyphrases
- trust region
- reinforcement learning
- global optimum
- optimization methods
- column generation
- objective function
- newton method
- log likelihood
- function approximation
- line search
- state space
- mean shift
- learning algorithm
- machine learning
- hessian matrix
- levenberg marquardt
- neural network
- dynamic programming
- artificial neural networks
- linear equations
- simulated annealing
- search space