Accuracy (Cross-Subject): 84.8Accuracy (Cross-Setup): 86.2GFLOPS per prediction: 32.4
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): 84.5Accuracy (Cross-Setup): 86.3
Accuracy (Cross-Subject): 86.2Accuracy (Cross-Setup): 88.4
Accuracy (Cross-Subject): 85.9Accuracy (Cross-Setup): 84.3
Accuracy (Cross-Subject): 86.6%Accuracy (Cross-Setup): 89.0 %
Accuracy (Cross-Subject): 85.4%Accuracy (Cross-Setup): 87.4%
Accuracy (Cross-Subject): 70.8%Accuracy (Cross-Setup): 71.6%
Accuracy (Cross-Subject): 68.3%Accuracy (Cross-Setup): 67.2%
Accuracy (Cross-Subject): 80.5%Accuracy (Cross-Setup): 83.2%
Accuracy (Cross-Subject): 81.1%Accuracy (Cross-Setup): 82.7%
Skeleton Image Representation for 3D Action Recognition based on Tree Structure and Reference Joints
Accuracy (Cross-Subject): 67.9%Accuracy (Cross-Setup): 62.8%
Accuracy (Cross-Subject): 82.7%Accuracy (Cross-Setup): 84.7%
Accuracy (Cross-Subject): 85.9%Accuracy (Cross-Setup): 87.6%Ensembled Modalities: 4
USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelation
Accuracy (Cross-Subject): 80.6Accuracy (Cross-Setup): 79.3
Accuracy (Cross-Subject): 78.3%Accuracy (Cross-Setup): 79.8%
Accuracy (Cross-Subject): 86.3Accuracy (Cross-Setup): 87.8