Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Nanjing) Co., Ltd
1
0
3
A single neural reranker can outperform specialized models in diverse environments while being highly sample-efficient, achieving 93% of full performance with just 9.2% of the target-domain data.