Optimizing Retrieval for RAG via Reinforced Contrastive Learning

Jiawei Zhou, Lei Chen · 2025

Read Optimizing Retrieval for RAG via Reinforced Contrastive Learning by Jiawei Zhou, Lei Chen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-24652.txt.

Description

Jiawei Zhou and Lei Chen aim to make search work better when its user is an AI, not a person. In such setups, what counts as a good hit is hard to set in advance. Old search tools learn from fixed human labels, so they adapt poorly to new tasks.

Their method, R3, trains the search tool by trial and reward inside the task setup itself. The tool can then test new moves and improve with little hand tuning. On a range of tasks, R3 raised end scores by more than five points over the base tool. It also beat strong past tools.

Catalog record

CREATORJiawei Zhou, Lei Chen
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE88 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-24652

Browse the full library · About the Tower of Babel