ERank: Fusing Supervised Fine-Tuning and Reinforcement Learning for Effective and Efficient Text Reranking
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (36) · 30121-30129
DOI: 10.1609/aaai.v40i36.40261 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Yuzheng Cai und 5 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref).
Abstract (Originalsprache)
Text reranking models are a crucial component in modern systems like Retrieval-Augmented Generation, tasked with selecting the most relevant documents prior to generation. However, current Large Language Models (LLMs) powered rerankers often face a fundamental trade-off. On one hand, Supervised Fine-Tuning based pointwise methods that frame relevance as a binary classification task lack the necessary scoring discrimination, particularly for those built on reasoning LLMs. On the other hand, approaches designed for complex reasoning often employ powerful yet inefficient listwise formulations, rendering them impractical for low latency applications. To resolve this dilemma, we introduce ERank, a highly Effective and Efficient pointwise reranker built from a reasoning LLM that excels across diverse relevance scenarios. We propose a novel two-stage training pipeline that begins with Supervised Fine-Tuning (SFT). In this stage, we move beyond binary labels and train the model generatively to output fine grained integer scores, which significantly enhances relevance discrimination. The model is then further refined using Reinforcement Learning (RL) with a novel, listwise derived reward. This technique instills global ranking awareness into the efficient pointwise architecture. We evaluate the ERank reranker on the BRIGHT, FollowIR, TREC DL, and BEIR benchmarks, demonstrating superior effectiveness and robustness compared to existing approaches. On the reasoning-intensive BRIGHT benchmark, our ERank-4B achieves an nDCG@10 of 38.7, while a larger 32B variant reaches a state of the art nDCG@10 of 40.2.
Zitieren
APA
Cai, Y., Zhang, Y., Long, D., Li, M., Xie, P., & Zheng, W. (2026). ERank: Fusing Supervised Fine-Tuning and Reinforcement Learning for Effective and Efficient Text Reranking. Proceedings of the AAAI Conference on Artificial Intelligence, 40(36), 30121-30129. https://doi.org/10.1609/aaai.v40i36.40261
Vancouver
Cai Y, Zhang Y, Long D, Li M, Xie P, Zheng W. ERank: Fusing Supervised Fine-Tuning and Reinforcement Learning for Effective and Efficient Text Reranking. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(36):30121-30129. doi:10.1609/aaai.v40i36.40261
BibTeX
@article{Cai2026,
author = {Cai, Yuzheng and Zhang, Yanzhao and Long, Dingkun and Li, Mingxin and Xie, Pengjun and Zheng, Weiguo},
title = {ERank: Fusing Supervised Fine-Tuning and Reinforcement Learning for Effective and Efficient Text Reranking},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {36},
pages = {30121--30129},
doi = {10.1609/aaai.v40i36.40261}
}
EUVIMED