Elucidating the Design Space of Dataset Condensation

arXiv:2404.13733v1 Announce Type: new
Abstract: Dataset condensation, a concept within data-centric learning, efficiently transfers critical attributes from an original dataset to a synthetic version, maintaining both diversity and realism. This approach significantly improves model training efficiency and is adaptable across multiple application areas. Previous methods in dataset condensation have faced challenges: some incur high computational costs which limit scalability to larger datasets (e.g., MTT, DREAM, and TESLA), while others are restricted to less optimal design spaces, which could hinder potential …

abstract application arxiv challenges concept cs.ai cs.cv cs.lg data data-centric dataset design diversity efficiency multiple space synthetic training type