Antoni Wuczyński

Randomness underpins cryptography, simulation and machine learning, yet its quality varies widely across families. Judging that quality is harder than it looks, since the verdict depends on the test battery, the sample size and the extractor, and no single tool takes a source from generation through extraction to testing in one place. This dissertation asks whether quantum generators produce measurably better randomness than classical alternatives, and builds a unified framework to answer it.