Doubly Reparameterized Gradient Estimators for Monte Carlo Objectives.
George TuckerDieterich LawsonShixiang GuChris J. MaddisonPublished in: ICLR (Poster) (2019)
Keyphrases
- monte carlo
- gradient estimators
- importance sampling
- monte carlo simulation
- markov chain
- monte carlo methods
- adaptive sampling
- simulation study
- monte carlo tree search
- particle filter
- temporal difference
- matrix inversion
- optimal strategy
- point processes
- branch and bound
- variance reduction
- monte carlo method
- global illumination
- policy evaluation
- probabilistic model