Real-Fake: Effective Training Data Synthesis Through Distribution Matching

arXiv:2310.10402v2 Announce Type: replace
Abstract: Synthetic training data has gained prominence in numerous learning tasks and scenarios, offering advantages such as dataset augmentation, generalization evaluation, and privacy preservation. Despite these benefits, the efficiency of synthetic data generated by current methodologies remains inferior when training advanced deep models exclusively, limiting its practical utility. To address this challenge, we analyze the principles underlying training data synthesis for supervised learning and elucidate a principled theoretical framework from the distribution-matching perspective that explicates the …

abstract advanced advantages arxiv augmentation benefits cs.ai cs.lg current data dataset distribution efficiency evaluation fake generated practical preservation privacy synthesis synthetic synthetic data tasks through training training data type utility