io.improbable.keanu.vertices.dbl.nonprobabilistic.diff

Class AutoDiffBroadcast

java.lang.Object

io.improbable.keanu.vertices.dbl.nonprobabilistic.diff.AutoDiffBroadcast

public class AutoDiffBroadcast
extends java.lang.Object

This class is meant to help with auto diff in operations that support implicit broadcasting. E.g. In addition/subtraction/multiplication/division scalar operands can be operated with non-scalar operands.

Constructor Summary

Constructors

Constructor and Description
AutoDiffBroadcast()

Method Summary

Modifier and Type	Method and Description
static PartialDerivative	correctForBroadcastPartialForward(PartialDerivative partial, long[] partialOfShape, long[] targetOfShape)
static PartialDerivative	correctForBroadcastPartialReverse(PartialDerivative partial, long[] partialWrtShape, long[] targetWrtShape)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

AutoDiffBroadcast

public AutoDiffBroadcast()

Method Detail

correctForBroadcastPartialForward

public static PartialDerivative correctForBroadcastPartialForward(PartialDerivative partial,
                                                                  long[] partialOfShape,
                                                                  long[] targetOfShape)

correctForBroadcastPartialReverse

public static PartialDerivative correctForBroadcastPartialReverse(PartialDerivative partial,
                                                                  long[] partialWrtShape,
                                                                  long[] targetWrtShape)