Updaters/Optimizers
Special algorithms for gradient descent.

What are updaters?

The main difference among the updaters is how they treat the learning rate. Stochastic Gradient Descent, the most common learning algorithm in deep learning, relies on Theta (the weights in hidden layers) and alpha (the learning rate). Different updaters help optimize the learning rate until the neural network converges on its most performant state.

Usage

To use the updaters, pass a new class to the updater() method in either a ComputationGraph or MultiLayerNetwork.
1
ComputationGraphConfiguration conf = new NeuralNetConfiguration.Builder()
2
.updater(new Adam(0.01))
3
// add your layers and hyperparameters below
4
.build();
Copied!

Available updaters

NadamUpdater

[source]
applyUpdater
1
public void applyUpdater(INDArray gradient, int iteration, int epoch)
Copied!
Calculate the update based on the given gradient
    param gradient the gradient to get the update for
    param iteration
    return the gradient

NesterovsUpdater

[source]
Nesterov’s momentum. Keep track of the previous layer’s gradient and use it as a way of updating the gradient.
applyUpdater
1
public void applyUpdater(INDArray gradient, int iteration, int epoch)
Copied!
Get the nesterov update
    param gradient the gradient to get the update for
    param iteration
    return

RmsPropUpdater

[source]
RMS Prop updates:

AdaGradUpdater

[source]
Vectorized Learning Rate used per Connection Weight
applyUpdater
1
public void applyUpdater(INDArray gradient, int iteration, int epoch)
Copied!
Gets feature specific learning rates Adagrad keeps a history of gradients being passed in. Note that each gradient passed in becomes adapted over time, hence the opName adagrad
    param gradient the gradient to get learning rates for
    param iteration

AdaMaxUpdater

[source]
The AdaMax updater, a variant of Adam. http://arxiv.org/abs/1412.6980
applyUpdater
1
public void applyUpdater(INDArray gradient, int iteration, int epoch)
Copied!
Calculate the update based on the given gradient
    param gradient the gradient to get the update for
    param iteration
    return the gradient

NoOpUpdater

[source]
NoOp updater: gradient updater that makes no changes to the gradient

AdamUpdater

[source]
The Adam updater. http://arxiv.org/abs/1412.6980
applyUpdater
1
public void applyUpdater(INDArray gradient, int iteration, int epoch)
Copied!
Calculate the update based on the given gradient
    param gradient the gradient to get the update for
    param iteration
    return the gradient

AdaDeltaUpdater

[source]
Ada delta updater. More robust adagrad that keeps track of a moving window average of the gradient rather than the every decaying learning rates of adagrad
applyUpdater
1
public void applyUpdater(INDArray gradient, int iteration, int epoch)
Copied!
Get the updated gradient for the given gradient and also update the state of ada delta.
    param gradient the gradient to get the updated gradient for
    param iteration
    return the update gradient

SgdUpdater

[source]
SGD updater applies a learning rate only

GradientUpdater

[source]
Gradient modifications: Calculates an update and tracks related information for gradient changes over time for handling updates.

AMSGradUpdater

[source]
The AMSGrad updater Reference: On the Convergence of Adam and Beyond - https://openreview.net/forum?id=ryQu7f-RZ