Speaker
Mariia Baidachna
Description
Quantum machine learning faces several bottlenecks that impede its empirical and theoretical computational advantage. One critical challenge is encoding classical data to quantum states. In this work, we present a novel data augmentation strategy applied after encoding, resulting in faster convergence with less data on quantum machine learning models. We demonstrate its effectiveness on generative diffusion-inspired models, showing that even limited datasets can be utilized for learning distributions.
