Accuracy (Body + Fingers joints): 91.78Accuracy (Body joints): 89.56Accuracy (Body + Fingers + Face joints): 89.64
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 (Cross-Subject): 57.01 %
Accuracy: 93.96
Accuracy: 93.4
Accuracy (Cross-Subject): 90.3Accuracy (Cross-Setup): 91.5Ensembled Modalities: 4
Accuracy: 96.00%
Accuracy: 95.7%
1:3 Accuracy: 92.91:1 Accuracy: 95.933:1 Accuracy: 96.76Cross-person Accuracy: 88.70
1:1 Accuracy: 96.17
Accuracy: 96.00%
Accuracy: 90.9%
1:1 Accuracy: 94.43
Accuracy (Cross-Subject): 90.9Accuracy (Cross-Setup): 91.9
Accuracy (Cross-Subject): 90.7Accuracy (Cross-Setup): 91.8Ensembled Modalities: 6
Accuracy: 98.60%
Accuracy (Cross-Subject): 90.6Accuracy (Cross-Setup): 92.2Ensembled Modalities: 6
Accuracy (Cross-Subject): 90.3Accuracy (Cross-Setup): 91.7Ensembled Modalities: 6
Accuracy: 99.02%
Accuracy (Cross-Subject): 90.9Accuracy (Cross-Setup): 92.2Ensembled Modalities: 6
Accuracy (Cross-Subject): 90.4Accuracy (Cross-Setup): 91.7Ensembled Modalities: 4
Accuracy: 94.0%
Accuracy: 93.3%
Accuracy (Cross-Subject): 90.2Accuracy (Cross-Setup): 91.7Ensembled Modalities: 4
Accuracy: 97.9
Accuracy: 98.3%
Accuracy: 97.5