Safe Exploration in Markov Decision Processes with Time-Variant Safety using Spatio-Temporal Gaussian Process.
Akifumi WachiHiroshi KajinoAsim MunawarPublished in: CoRR (2018)
Keyphrases
- gaussian process
- markov decision processes
- spatio temporal
- model based reinforcement learning
- gaussian processes
- interval estimation
- optimal policy
- transition matrices
- reinforcement learning
- approximate inference
- dynamic programming
- regression model
- state space
- model selection
- bayesian framework
- latent variables
- semi supervised
- image sequences
- hyperparameters
- policy iteration
- decision theoretic planning
- partially observable
- action selection
- average reward
- moving objects
- human motion
- reward function
- markov decision process
- pairwise
- data mining