behavior cloning

GPTKB entity

Statements (23)
Predicate Object
gptkbp:instanceOf machine learning technique
gptkbp:alternativeTo inverse reinforcement learning
gptkbp:compatibleWith reward function
gptkbp:firstDescribed 1989
https://www.w3.org/2000/01/rdf-schema#label behavior cloning
gptkbp:improves data augmentation
gptkbp:input state-action pairs
gptkbp:learnsFrom demonstrations
gptkbp:limitation compounding errors
covariate shift
gptkbp:notableFor self-driving cars
robot manipulation
video game AI
gptkbp:output gptkb:public_policy
gptkbp:relatedTo gptkb:reinforcement_learning
supervised learning
gptkbp:requires expert demonstrations
gptkbp:trainer minimizing prediction error
gptkbp:usedIn robotics
autonomous driving
imitation learning
gptkbp:bfsParent gptkb:DAgger
gptkbp:bfsLayer 8