E^2Rank: Your Text Embedding can Also be an Effective and Efficient Listwise Reranker
Qi Liu, Yanzhao Zhang, Mingxin Li, Dingkun Long, Pengjun Xie, Jiaxin Mao
2025-10-28
Summary
This paper introduces a new method, E^2Rank, for improving how search engines rank results. It aims to make the initial search and the final re-ranking of those results more accurate and efficient using a single text embedding model.
What's the problem?
Current search systems use text embedding models to quickly find relevant documents. However, these models aren't as good at precisely ranking those documents as more complex systems, especially newer ones based on large language models. These better ranking systems are slower and require more computing power, creating a trade-off between speed and accuracy.
What's the solution?
E^2Rank solves this by taking a standard text embedding model and continuing to train it, but this time focusing on ranking documents as a whole list rather than individually. It does this by feeding the model the original search query *along with* the top potential results, essentially giving it more context. This allows the model to learn how documents relate to each other and to the query, improving ranking without sacrificing the speed of using embeddings.
Why it matters?
This research is important because it shows you can get both fast search results *and* highly accurate ranking with a single model. This means search engines can be more effective without needing significantly more computing resources, which is a big step forward for making information retrieval more accessible and efficient.
Abstract
Text embedding models serve as a fundamental component in real-world search applications. By mapping queries and documents into a shared embedding space, they deliver competitive retrieval performance with high efficiency. However, their ranking fidelity remains limited compared to dedicated rerankers, especially recent LLM-based listwise rerankers, which capture fine-grained query-document and document-document interactions. In this paper, we propose a simple yet effective unified framework E^2Rank, means Efficient Embedding-based Ranking (also means Embedding-to-Rank), which extends a single text embedding model to perform both high-quality retrieval and listwise reranking through continued training under a listwise ranking objective, thereby achieving strong effectiveness with remarkable efficiency. By applying cosine similarity between the query and document embeddings as a unified ranking function, the listwise ranking prompt, which is constructed from the original query and its candidate documents, serves as an enhanced query enriched with signals from the top-K documents, akin to pseudo-relevance feedback (PRF) in traditional retrieval models. This design preserves the efficiency and representational quality of the base embedding model while significantly improving its reranking performance. Empirically, E^2Rank achieves state-of-the-art results on the BEIR reranking benchmark and demonstrates competitive performance on the reasoning-intensive BRIGHT benchmark, with very low reranking latency. We also show that the ranking training process improves embedding performance on the MTEB benchmark. Our findings indicate that a single embedding model can effectively unify retrieval and reranking, offering both computational efficiency and competitive ranking accuracy.