no code implementations • ECCV 2020 • Mingeun Kang, Kiwon Lee, Yong H. Lee, Changho Suh
We consider graph-based semi-supervised learning that leverages a similarity graph across data points to better exploit data structure exposed in unlabeled data.