Can I read ClusterE-ZSL A Novel Cluster-Based Embedding For Enhanced Zero-Shot Learning in Contrastive Pre-Training Cross-Modal Retrieval on EtoBox?

ClusterE-ZSL A Novel Cluster-Based Embedding For Enhanced Zero-Shot Learning in Contrastive Pre-Training Cross-Modal Retrieval by Kamran Aziz is a document available to read on EtoBox.

What is ClusterE-ZSL A Novel Cluster-Based Embedding For Enhanced Zero-Shot Learning in Contrastive Pre-Training Cross-Modal Retrieval about?

The document presents ClusterE-ZSL, a novel cluster-based embedding approach designed to enhance zero-shot learning (ZSL) for cross-modal retrieval tasks. By utilizing an iterative training strategy and cross-modal contrastive learning, the model achieves significant improvements in accuracy on benchmark datasets like Flickr30K and MSCOCO, outperforming existing methods. The proposed framework effectively addresses the semantic gap between text and image modalities, improving the scalability and efficiency

Author

Kamran Aziz

Language

EN