Statements (26)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
neural network technique
|
| gptkbp:appliesTo |
actor-critic methods
Deep Q-Networks |
| gptkbp:hasConcept |
add parameterized noise to neural network weights
|
| https://www.w3.org/2000/01/rdf-schema#label |
Noisy Nets
|
| gptkbp:organization |
gptkb:DeepMind
|
| gptkbp:proposedBy |
gptkb:Demis_Hassabis
gptkb:Alex_Graves gptkb:Bilal_Piot Charles Blundell Ian Osband Jacob Menick Meire Fortunato Mohammad Gheshlaghi Azar Olivier Pietquin Remi Munos Vlad Mnih |
| gptkbp:publicationYear |
2017
|
| gptkbp:publishedIn |
arXiv:1706.10295
|
| gptkbp:purpose |
improve exploration in reinforcement learning
|
| gptkbp:relatedTo |
exploration-exploitation tradeoff
parameter noise randomized value functions |
| gptkbp:usedIn |
gptkb:reinforcement_learning
|
| gptkbp:bfsParent |
gptkb:Rainbow_DQN
|
| gptkbp:bfsLayer |
7
|