FREE: Faster and Better Data-Free Meta-Learning

arXiv:2405.00984v1 Announce Type: new
Abstract: Data-Free Meta-Learning (DFML) aims to extract knowledge from a collection of pre-trained models without requiring the original data, presenting practical benefits in contexts constrained by data privacy concerns. Current DFML methods primarily focus on the data recovery from these pre-trained models. However, they suffer from slow recovery speed and overlook gaps inherent in heterogeneous pre-trained models. In response to these challenges, we introduce the Faster and Better Data-Free Meta-Learning (FREE) framework, which contains: (i) a …

abstract arxiv benefits collection concerns cs.cv cs.lg current data data privacy data recovery extract faster focus free however knowledge meta meta-learning practical presenting pre-trained models privacy recovery speed type