gov.sandia.cognition.learning.function.cost

Interface ParallelizableCostFunction

All Superinterfaces:

java.lang.Cloneable, CloneableSerializable, CostFunction<Evaluator<? super Vector,? extends Vector>,java.util.Collection<? extends InputOutputPair<? extends Vector,Vector>>>, DifferentiableCostFunction, Evaluator<Evaluator<? super Vector,? extends Vector>,java.lang.Double>, PerformanceEvaluator<Evaluator<? super Vector,? extends Vector>,java.util.Collection<? extends InputOutputPair<? extends Vector,Vector>>,java.lang.Double>, java.io.Serializable, Summarizer<TargetEstimatePair<? extends Vector,? extends Vector>,java.lang.Double>, SupervisedCostFunction<Vector,Vector>, SupervisedPerformanceEvaluator<Vector,Vector,Vector,java.lang.Double>

All Known Implementing Classes:

AbstractParallelizableCostFunction, SumSquaredErrorCostFunction

public interface ParallelizableCostFunction
extends SupervisedCostFunction<Vector,Vector>, DifferentiableCostFunction

Interface describing a cost function that can (largely) be computed in parallel.

Since:

2.1

Author:

Kevin R. Dixon

Method Summary

Modifier and Type	Method and Description
Vector	computeParameterGradientAmalgamate(java.util.Collection<java.lang.Object> partialResults) Amalgamates the linear components of the cost gradient function into a single Vector.
java.lang.Object	computeParameterGradientPartial(GradientDescendable function) Computes the partial (linear) component of the cost function gradient.
java.lang.Double	evaluateAmalgamate(java.util.Collection<java.lang.Object> partialResults) Amalgamates the linear components of the cost function into a single Double.
java.lang.Object	evaluatePartial(Evaluator<? super Vector,? extends Vector> evaluator) Computes the partial (linear) component of the cost function.

Methods inherited from interface gov.sandia.cognition.learning.function.cost.DifferentiableCostFunction

computeParameterGradient

Methods inherited from interface gov.sandia.cognition.learning.function.cost.CostFunction

clone, evaluate, getCostParameters, setCostParameters

Methods inherited from interface gov.sandia.cognition.util.Summarizer

summarize

Methods inherited from interface gov.sandia.cognition.learning.performance.SupervisedPerformanceEvaluator

evaluatePerformance

Methods inherited from interface gov.sandia.cognition.learning.performance.PerformanceEvaluator

evaluatePerformance

Method Detail

evaluatePartial

java.lang.Object evaluatePartial(Evaluator<? super Vector,? extends Vector> evaluator)

Computes the partial (linear) component of the cost function. This portion will be performed in parallel.

Parameters:

evaluator - Evaluator to compute the cost of

Returns:

Object that contains the linear component of the cost function

evaluateAmalgamate

java.lang.Double evaluateAmalgamate(java.util.Collection<java.lang.Object> partialResults)

Amalgamates the linear components of the cost function into a single Double. This portion will be performed in sequence.

Parameters:

partialResults - Collection of partial (linear) results

Returns:

Cost function of the partial results

computeParameterGradientPartial

java.lang.Object computeParameterGradientPartial(GradientDescendable function)

Computes the partial (linear) component of the cost function gradient. This portion will be performed in parallel.

Parameters:

function - GradientDescendable to compute the gradient of

Returns:

Object that contains the linear component of the gradient

computeParameterGradientAmalgamate

Vector computeParameterGradientAmalgamate(java.util.Collection<java.lang.Object> partialResults)

Amalgamates the linear components of the cost gradient function into a single Vector. This portion will be performed in sequence.

Parameters:

partialResults - Collection of partial (linear) gradient components

Returns:

Vector describing the gradient