Accuracy (CV): 96.1Accuracy (CS): 92.1
Papers with Code · CC BY-SA 4.0
Zero-Shot Learning
Zero-shot learning (ZSL) is a model's ability to detect classes never seen during training. The condition is that the classes are not known during supervised learning.
Earlier work in zero-shot learning use attributes in a two-step approach to infer unknown classes. In the computer vision context, more recent advances learn mappings from image feature space to semantic space. Other approaches learn non-linear multimodal embeddings. In the modern NLP context, language models can be evaluated on downstream tasks without fine tuning.
Benchmark datasets for zero-shot learning include aPY, AwA, and CUB, among others.
( Image credit: Prototypical Networks for Few shot Learning in PyTorch )
Further readings:
논문
Accuracy (CV): 97.4Accuracy (CS): 92.6Ensembled Modalities: 4
Accuracy (CV): 96.5Accuracy (CS): 90.7
Accuracy (CV): 96.3Accuracy (CS): 90
Accuracy (CV): 94.90Accuracy (CS): 89.57Ensembled Modalities: Only joint data
Accuracy (CV): 96.8Accuracy (CS): 92.4Ensembled Modalities: 4
Accuracy (CV): 96.4Accuracy (CS): 91.5
Accuracy (CV): 96.2Accuracy (CS): 91.5
Accuracy (CV): 96.0Accuracy (CS): 91.5
Accuracy (CV): 96.3Accuracy (CS): 90.2
Accuracy (CV): 96.3Accuracy (CS): 90.3
Accuracy (CV): 96.4Accuracy (CS): 91.7
Accuracy (CV): 96.6Accuracy (CS): 92.0
Accuracy (CV): 96.7Accuracy (CS): 91.6Ensembled Modalities: 4
Accuracy (CV): 96Accuracy (CS): 89.7
Accuracy (CV): 97.8Accuracy (CS): 92.6
Accuracy (CV): 96.1Accuracy (CS): 89.9
Accuracy (CV): 96.1Accuracy (CS): 89.9
Accuracy (CV): 96.2Accuracy (CS): 90.0
Accuracy (CV): 96.5Accuracy (CS): 90.7Ensembled Modalities: 4
Accuracy (CV): 95.74Accuracy (CS): 89.58
Accuracy (CV): 96Accuracy (CS): 90.9
Symbiotic Graph Neural Networks for 3D Skeleton-based Human Action Recognition and Motion Prediction
Accuracy (CV): 96.4Accuracy (CS): 90.1
Accuracy (CV): 96.5Accuracy (CS): 91.0
Accuracy (CV): 94.75Accuracy (CS): 91.56
Accuracy (CV): 96.4Accuracy (CS): 90.3