Novel Class Discovery

3 benchmarks65 papers

The goal of Novel Class Discovery (NCD) is to identify new classes in unlabeled data, by exploiting prior knowledge from known classes. In this specific setup, the data is split in two sets. The first is a labeled set containing known classes and the second is an unlabeled set containing unknown classes that must be discovered.

Benchmarks

Novel Class Discovery on SVHN

Clustering Accuracy

Novel Class Discovery on cifar10

Clustering Accuracy

Novel Class Discovery on cifar100

Clustering Accuracy