Add to Favourites
To login click here
such as textual descriptions, attributes, or class hierarchies.
Model Training: A model is trained on the seen classes, leveraging the auxiliary information to learn a generalised representation of the data.
Inference: The trained model is used to make predictions on unseen classes, using the learned knowledge to accurately classify novel instances.
Zero-Shot Learning is a powerful technique that enables machines to learn and generalise from previously unseen data with astonishing accuracy. By leveraging semantic relationships and attribute-based representations, ZSL is able to bridge the gap between seen and unseen classes, allowing machines to identify and classify novel instances without explicit training.