computer-vision-5
Generalized Zero-Shot Learning
3260 papers • 126 benchmarks • 313 datasets
In generalized zero shot learning (GZSL), the set of classes are split into seen and unseen classes, where training relies on the semantic features of the seen and unseen classes and the visual representations of only the seen classes, while testing uses the visual representations of the seen and unseen classes.
(Image credit: Papersgraph)
Benchmarks
These leaderboards are used to track progress in generalized-zero-shot-learning-5
Trend
Dataset
Best Model
Actions
SUN Attribute
SUN Attribute
AwA2
AwA2
CUB-200-2011
CUB-200-2011
Libraries
iUse these libraries to find generalized-zero-shot-learning-5 models and implementations
Datasets
Subtasks
No subtasks available.
Most implemented papers
Zero-Shot Learning with Common Sense Knowledge Graphs
Stephen H. Bach, Nihal V. Nayak•Wed Jun 17 2020
It is found that non-linear aggregators such as LSTMs or transformers lead to significant improvements on zero-shot tasks, and ZSL-KG outperforms the best performing graph neural networks with linear aggregators by an average of 3.8 points of accuracy.
39
