Wang, Lixu, Xinyu Du, and Qi Zhu. “Semantic feature learning for universal unsupervised cross-domain retrieval.” Advances in Neural Information Processing Systems 37 (2024): 79516-79539.
Paper Link: NeurIPS 2024
Introduction
This paper proposes semantic feature learning for universal unsupervised cross-domain retrieval. The approach focuses on learning semantic representations that are effective across diverse domains.
Key Contributions
- Universal cross-domain retrieval framework
- Semantic feature learning methodology
- Effective retrieval across multiple domain pairs
Methodology
The method learns semantic features that capture domain-invariant information, enabling universal retrieval performance across different domain combinations.
Figures
Note: Additional figures will be added once the paper is available on arXiv.
Results
The approach demonstrates strong performance in universal cross-domain retrieval scenarios, showing effectiveness across various domain pairs.