SOTAVerified

Novel Class Discovery

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.

Papers

Showing 1120 of 65 papers

TitleStatusHype
When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral AnalysisCode1
Class-relation Knowledge Distillation for Novel Class DiscoveryCode1
Open-world Semi-supervised Novel Class DiscoveryCode1
Large-scale Pre-trained Models are Surprisingly Strong in Incremental Novel Class DiscoveryCode1
Novel Class Discovery for 3D Point Cloud Semantic SegmentationCode1
On-the-Fly Category DiscoveryCode1
Bootstrap Your Own Prior: Towards Distribution-Agnostic Novel Class DiscoveryCode1
Parametric Classification for Generalized Category Discovery: A Baseline StudyCode1
Learning to Discover and Detect ObjectsCode1
Modeling Inter-Class and Intra-Class Constraints in Novel Class DiscoveryCode1
Show:102550
← PrevPage 2 of 7Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1AutoNovelClustering Accuracy0.92Unverified
#ModelMetricClaimedVerifiedStatus
1AutoNovelClustering Accuracy0.75Unverified
#ModelMetricClaimedVerifiedStatus
1AutoNovelClustering Accuracy0.95Unverified