Hybrid distributions for finite binary generators -- definitions #
The cut-th hybrid uses generator output in coordinates strictly below cut
and independent uniform bits in the remaining coordinates. Both sources of
randomness are packed into one Boolean string, so every probability remains an
exact eventProb over a finite dyadic sample space.
Complementing a test records the one-bit polarity choice used to orient an absolute distinguishing advantage before applying the hybrid argument.
The cut-th hybrid between uniform output bits and the generator output.
Coordinates below cut come from the generator; all later coordinates are
independent uniform bits.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Combined random strings whose cut-th hybrid output is accepted.
Equations
- generator.hybridAcceptedRandomness test cut = {randomness : Fin (seedLength + outputLength) → Bool | generator.hybridOutput cut randomness ∈ test}
Instances For
Acceptance probability of a test under the cut-th hybrid.
Equations
- generator.hybridAcceptanceProbability test cut = Complexity.eventProb (generator.hybridAcceptedRandomness test cut)
Instances For
Change in test acceptance when one more output coordinate is switched from uniform to generated.
Equations
- generator.hybridGap test step = generator.hybridAcceptanceProbability test (step + 1) - generator.hybridAcceptanceProbability test step