Statements (23)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
mathematical optimization
|
| gptkbp:address |
non-convergence of Adam
|
| gptkbp:advantage |
guaranteed convergence under certain conditions
|
| gptkbp:author |
gptkb:Sanjiv_Kumar
Sashank J. Reddi Satyen Kale |
| gptkbp:citation |
gptkb:ICLR
2018 On the Convergence of Adam and Beyond |
| gptkbp:designedFor |
gptkb:machine_learning
|
| https://www.w3.org/2000/01/rdf-schema#label |
AMSGrad
|
| gptkbp:implementedIn |
gptkb:TensorFlow
gptkb:PyTorch |
| gptkbp:introduced |
Sashank J. Reddi
|
| gptkbp:introducedIn |
2018
|
| gptkbp:publishedIn |
gptkb:ICLR_2018
|
| gptkbp:usedIn |
deep learning
neural network training |
| gptkbp:uses |
momentum
adaptive learning rates |
| gptkbp:variant |
gptkb:Adam
|
| gptkbp:bfsParent |
gptkb:Adaptive_Moment_Estimation
|
| gptkbp:bfsLayer |
6
|