bham.leakiest

Class ApproxPrior

java.lang.Object

bham.leakiest.ApproxPrior

public class ApproxPrior
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
ApproxPrior()

Method Summary

Modifier and Type	Method and Description
static double[]	errorMinEntropyLeakSmallProbsRemoved(ProbDist apd, Channel[] channels, double sumOfSmallProbs) Returns an error of the leakage caused by removing small probabilities by the input approximation technique.
static double[]	errorMinEntropyLeakSmallProbsRemovedNoReexecutionWithJointInput(ProbDist pd, Channel[] channels, double sumOfSmallProbs) Returns an error of the leakage caused by removing small probabilities by the input approximation technique in the case of jointly supported input distributions.
static double[]	errorMinEntropyLeakSmallProbsRemovedNoReexecutionWithSharedInput(ProbDist pd, Channel[] channels, double sumOfSmallProbs) Returns an error of the leakage caused by removing small probabilities by the input approximation technique in the case of shared input distributions.
static double	sumOfProbsRemoved(ProbDist apd) Returns the summation of all probabilities removed from the input distribution, by input approximation.

Methods inherited from class java.lang.Object

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

Constructor Detail

ApproxPrior

public ApproxPrior()

Method Detail

errorMinEntropyLeakSmallProbsRemoved

public static double[] errorMinEntropyLeakSmallProbsRemoved(ProbDist apd,
                                                            Channel[] channels,
                                                            double sumOfSmallProbs)

Returns an error of the leakage caused by removing small probabilities by the input approximation technique. This method is used when we esitmate a leakage measure of a huge composed channel. This error is valid if we can calculate the min-entropies of channels with the approximate input, and invalid if we cannot caluclate them.

Parameters:

apd - approximate input distribution

channels - all channels that might be composed

sumOfSmallProbs - the sum of small probabilities removed from the input distribution

Returns:

an error of the leakage calculated by input approximation

errorMinEntropyLeakSmallProbsRemovedNoReexecutionWithJointInput

public static double[] errorMinEntropyLeakSmallProbsRemovedNoReexecutionWithJointInput(ProbDist pd,
                                                                                       Channel[] channels,
                                                                                       double sumOfSmallProbs)

Returns an error of the leakage caused by removing small probabilities by the input approximation technique in the case of jointly supported input distributions. This method is used when we esitmate a leakage measure of a huge composed channel. This error is valid even if we cannot calculate the min-entropies of sub-channels while giving worse bounds.

Parameters:

pd - input distribution

channels - all channels that might be composed

sumOfSmallProbs - the sum of small probabilities removed from the input distribution

Returns:

an error of the leakage calculated by input approximation

errorMinEntropyLeakSmallProbsRemovedNoReexecutionWithSharedInput

public static double[] errorMinEntropyLeakSmallProbsRemovedNoReexecutionWithSharedInput(ProbDist pd,
                                                                                        Channel[] channels,
                                                                                        double sumOfSmallProbs)

Returns an error of the leakage caused by removing small probabilities by the input approximation technique in the case of shared input distributions. This method is used when we esitmate a leakage measure of a huge composed channel. This error is valid even if we cannot calculate the min-entropies of sub-channels while giving worse bounds.

Parameters:

pd - input distribution

channels - all channels that might be composed

sumOfSmallProbs - the sum of small probabilities removed from the input distribution

Returns:

an error of the leakage calculated by input approximation

sumOfProbsRemoved

public static double sumOfProbsRemoved(ProbDist apd)

Returns the summation of all probabilities removed from the input distribution, by input approximation.

Parameters:

apd - approximate input distribution

Returns:

the summation of all probabilities removed from the input distribution