Trust Region Policy Optimization (TRPO)

GPTKB entity

Predicate	Object
gptkbp:instanceOf	reinforcement learning algorithm
gptkbp:abbreviation	gptkb:TRPO
gptkbp:application	robotics game playing control tasks
gptkbp:author	John Schulman, Sergey Levine, Philipp Moritz, Michael Jordan, Pieter Abbeel
gptkbp:citation	gptkb:Trust_Region_Policy_Optimization gptkb:ICML 2015
gptkbp:designedFor	policy optimization
https://www.w3.org/2000/01/rdf-schema#label	Trust Region Policy Optimization (TRPO)
gptkbp:improves	policy gradient methods
gptkbp:influenced	gptkb:Soft_Actor-Critic_(SAC) gptkb:Proximal_Policy_Optimization_(PPO) modern policy gradient methods
gptkbp:input	policy network
gptkbp:introduced	gptkb:John_Schulman
gptkbp:introducedIn	2015
gptkbp:objective	maximize expected reward
gptkbp:openSource	gptkb:OpenAI_Baselines gptkb:Stable_Baselines
gptkbp:optimizedFor	constrained optimization
gptkbp:output	improved policy network
gptkbp:prohibits	trust region constraint
gptkbp:publishedIn	gptkb:ICML_2015
gptkbp:relatedTo	gptkb:Natural_Policy_Gradient gptkb:Proximal_Policy_Optimization_(PPO)
gptkbp:supportsAlgorithm	on-policy
gptkbp:type	KL-divergence constraint monotonic policy improvement
gptkbp:usedIn	deep reinforcement learning
gptkbp:bfsParent	gptkb:John_Schulman
gptkbp:bfsLayer	6