Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Open-world semi-supervised learning research benchmarks · SOTA2 Research

Open-world semi-supervised learning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

CIFAR-100 (test)

73.32Overall Accuracy

40

Feb 26, 2026

CIFAR-10 (test)

93.5Overall Accuracy

28

Feb 26, 2026

96.8Accuracy (Seen)

25

Feb 26, 2026

CIFAR-100 50% known, 50% novel

74.1Known Accuracy

16

Feb 26, 2026

89.1Seen Accuracy

16

Feb 26, 2026

66.9Seen Accuracy

16

Feb 26, 2026

89.9Seen Accuracy

13

Feb 26, 2026

Oxford-IIIT Pet 50% known / 50% novel classes (test)

95.3Accuracy (Known)

11

May 1, 2026

CIFAR-10 50% known, 50% novel

95.7Known Accuracy

9

Feb 26, 2026

ImageNet-100 50% known, 50% novel

89.6Known Acc

7

Feb 26, 2026

98.4Known Accuracy

6

May 1, 2026

95.5Known Accuracy

6

May 1, 2026

83.7Known Accuracy

6

May 1, 2026

Tiny ImageNet 50% known 50% novel classes (test)

58.3Known Accuracy

5

Feb 26, 2026

Tiny ImageNet (test)

59.1Seen Accuracy

4

Feb 26, 2026

ImageNet 100 25/75 seen/novel split (unlabelled)

Dual Ranking Statistics

89.5Seen Accuracy

4

Feb 26, 2026

ImageNet-100 50 50 seen novel split (unlabelled)

89.1Seen Accuracy

4

Feb 26, 2026

Popular datasets

ImageNet CIFAR CUB CIFAR-100 Tiny-ImageNet

Follow for updates

Join SOTA2 Contact