L2Regularization

class L2Regularization[T](val regularizationConstant: Double, stepSize: Double, maxIter: Int, tolerance: Double, minImprovementWindow: Int)(implicit vspace: MutableFiniteCoordinateField[T, _, Double], rand: RandBasis) extends StochasticGradientDescent[T]

Implements the L2 regularization update.

Each step is:

x_{t+1}i = (s_{ti} * x_{ti} - \eta * g_ti) / (eta * regularization + delta + s_ti)

where g_ti is the gradient and s_ti = \sqrt(\sum_t'^{t} g_ti^2)

trait Serializable
class Object
trait Matchable
class Any

Type members

Classlikes

case class History(sumOfSquaredGradients: T)

Inherited types

type State = State[T, Info, History]
Inherited from:
FirstOrderMinimizer

Value members

Concrete methods

override def initialHistory(f: StochasticDiffFunction[T], init: T): History
Definition Classes
override def updateHistory(newX: T, newGrad: T, newValue: Double, f: StochasticDiffFunction[T], oldState: State): History
Definition Classes

Inherited methods

protected def logger: LazyLogger
Inherited from:
SerializableLogging
def minimize(f: StochasticDiffFunction[T], init: T): T
Inherited from:
FirstOrderMinimizer

Concrete fields