AUC: 0.616f-Score: 0.081Intersection over Union: 0.052
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
edit success: 96.74portability: 60.41locality: 92.87fluency: 586.34
AUC: 0.765f-Score: 0.361Intersection over Union: 0.249
AUC: 0.542f-Score: 0.233Intersection over Union: 0.139
AUC: .839Balanced Accuracy: .770
AUC: 0.839f-Score: 0.341Intersection over Union: 0.233
AUC: .553Balanced Accuracy: .391
Accuracy: 76.9%
Accuracy: 85.7%
AUC: .733Balanced Accuracy: .514
Average accuracy of 3 splits: 89.22
Accuracy: 75.5%
AUC: .680Balanced Accuracy: .635
Accuracy: 83.14%
Accuracy (RGB+pose): -Accuracy (pose): 77.2
Accuracy: 77.9%
Accuracy: 86.9%
Accuracy (RGB+pose): -Accuracy (pose): 65.5
Accuracy: 73.4%
AUC: .770Balanced Accuracy: .680
Accuracy: 70.9
Accuracy: 86.7%
Accuracy: 77.5%