Accuracy (CV): 88.84Accuracy (CS): 83.36
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): 89.5Accuracy (CS): 82.0
Accuracy (CV): 84.6Accuracy (CS): 79.6
Accuracy (CV): 84.8Accuracy (CS): 79.6
Accuracy (CV): 88.8Accuracy (CS): 84.3
Accuracy (CV): 87.1Accuracy (CS): 79.8
Accuracy (CV): 93.2Accuracy (CS): 85.1
Accuracy (CV): 89.0Accuracy (CS): 81.8
Accuracy (CV): 89.27Accuracy (CS): 84.23
Accuracy (CV): 92.6Accuracy (CS): 84.1
Accuracy (CV): 88.4Accuracy (CS): 80.7
Accuracy (CV): 87.2Accuracy (CS): 80.0Ensembled Modalities: 4
Accuracy (CV): 84.00Accuracy (CS): 76.10
Accuracy (CV): 88.0Accuracy (CS): 81.8
Accuracy (CV): 86.1Accuracy (CS): 80.9
Accuracy (CV): 90.05Accuracy (CS): 82.83
Accuracy (CV): 93.22Accuracy (CS): 82.67
Accuracy (CV): 91.7Accuracy (CS): 85.2
Accuracy (CV): 93.4Accuracy (CS): 85.6
Accuracy (CV): 88.8Accuracy (CS): 80.7
Accuracy (CV): 93.0Accuracy (CS): 85.8
Accuracy (CV): 84.7Accuracy (CS): 76.5
Accuracy (CS): 85.0
Accuracy (CV): 89.3Accuracy (CS): 83.2
Accuracy (CV): 92.4Accuracy (CS): 84.8
Accuracy (CV): 81.2Accuracy (CS): 75.9
Accuracy (CV): 88.3Accuracy (CS): 81.5
Accuracy (CV): 92.8Accuracy (CS): 85.3
Accuracy (CV): 87.6Accuracy (CS): 79.2