Beyond Single Embeddings: Capturing Diverse Targets with Multi-Query Retrieval
Published in Arxiv, 2025
This paper introduces a new retriever architecture, Autoregressive Multi-Embedding Retriever (AMER), which autoregressively generates multiple query vectors, and all the predicted query vectors are used to retrieve documents from the corpus. We demonstrate the potential of using a multi-query vector retriever and open up a new direction for future work.
Recommended citation: Hung-Ting Chen, Xiang Liu , Shauli Ravfogel, Eunsol Choi. (2025). "Beyond Single Embeddings: Capturing Diverse Targets with Multi-Query Retrieval." Arxiv.
Download Paper