Accuracy: 89.3%
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: 91.1%
Accuracy (Cross-Subject): 25.5%Accuracy (Cross-Setup): 26.3%
Accuracy (Cross-Subject): 67.7%Accuracy (Cross-Setup): 66.9%
Accuracy (Cross-Subject): 58.4%Accuracy (Cross-Setup): 57.9%
Accuracy: 96.0
Accuracy: 92.0
mAP@0.50 (CV): 94.2mAP@0.50 (CS): 92.6
Accuracy (Cross-Subject): 36.3%Accuracy (Cross-Setup): 44.9%
Accuracy (Cross-Subject): 60.3%Accuracy (Cross-Setup): 63.2%
Accuracy (Cross-Subject): 58.3%Accuracy (Cross-Setup): 59.2%
Accuracy (Cross-Subject): 50.8%Accuracy (Cross-Setup): 54.7%
Accuracy (Cross-Subject): 62.2%Accuracy (Cross-Setup): 61.8%
Accuracy: 86.0%
mAP@0.50 (CV): 94.4mAP@0.50 (CS): 92.9
Accuracy: 69.1%
Accuracy (Cross-Subject): 64.6%Accuracy (Cross-Setup): 66.9%
Accuracy: 90.94
Skeleton Image Representation for 3D Action Recognition based on Tree Structure and Reference Joints
Accuracy (Cross-Subject): 65.5%Accuracy (Cross-Setup): 59.7%
Accuracy (Cross-Subject): 58.2%Accuracy (Cross-Setup): 60.9%
mAP@0.50 (CV): 93.7mAP@0.50 (CS): 90.4
Accuracy (Cross-Subject): 61.2%Accuracy (Cross-Setup): 63.3%
Accuracy (Cross-Subject): 59.9%Accuracy (Cross-Setup): 62.4%
Accuracy (Cross-Subject): 55.7%Accuracy (Cross-Setup): 57.9%
Accuracy: 85.4%
Accuracy: 92.91%