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

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 51–65 of 65 papers

Title	Date	Tasks	Status	Hype
Novel Class Discovery without Forgetting	Jul 21, 2022	Novel Class Discovery	—Unverified	0
Class-incremental Novel Class Discovery	Jul 18, 2022	Incremental LearningKnowledge Distillation	CodeCode Available	1
Towards Realistic Semi-Supervised Learning	Jul 5, 2022	Novel Class DiscoveryOpen-World Semi-Supervised Learning	CodeCode Available	1
Mutual Information-guided Knowledge Transfer for Novel Class Discovery	Jun 24, 2022	Novel Class DiscoveryRelation	—Unverified	0
Spacing Loss for Discovering Novel Categories	Apr 22, 2022	Novel Class Discovery	CodeCode Available	2
Open Set Domain Adaptation By Novel Class Discovery	Mar 7, 2022	Domain AdaptationNovel Class Discovery	—Unverified	0
Divide and Conquer: Compositional Experts for Generalized Novel Class Discovery	Jan 1, 2022	Novel Class Discovery	CodeCode Available	1
Novel Class Discovery in Semantic Segmentation	Dec 3, 2021	image-classificationImage Classification	CodeCode Available	1
A Unified Objective for Novel Class Discovery	Aug 19, 2021	Novel Class DiscoveryNovel Object Detection	CodeCode Available	1
AutoNovel: Automatically Discovering and Learning Novel Visual Categories	Jun 29, 2021	ClusteringImage Clustering	CodeCode Available	1
Open-Set Representation Learning through Combinatorial Embedding	Jun 29, 2021	Image CategorizationImage Retrieval	—Unverified	0
Neighborhood Contrastive Learning for Novel Class Discovery	Jun 20, 2021	ClusteringContrastive Learning	CodeCode Available	1
Meta Discovery: Learning to Discover Novel Classes given Very Limited Data	Feb 8, 2021	ClusteringMeta-Learning	CodeCode Available	1
Encoding Structure-Texture Relation with P-Net for Anomaly Detection in Retinal Images	Aug 9, 2020	AnatomyAnomaly Detection	CodeCode Available	1
OpenMix: Reviving Known Knowledge for Discovering Novel Visual Categories in An Open World	Apr 12, 2020	ClusteringNovel Class Discovery	—Unverified	0

Show:10 25 50

← PrevPage 3 of 3Next →

All datasets CIFAR10 cifar100 SVHN

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	AutoNovel	Clustering Accuracy	0.92	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	AutoNovel	Clustering Accuracy	0.75	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	AutoNovel	Clustering Accuracy	0.95	—	Unverified