DSL-R1: From SQL to DSL for Training Retrieval Agents across Structured and Unstructured Data with Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hu, Yunhai, Zhou, Junwei, Cao, Yumo, Long, Yitao, Xu, Yiwei, Jiang, Qiyi, Wang, Weiyao, Cao, Xiaoyu, Sun, Zhen, Zou, Yiran, Du, Nan
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911534864662528
author Hu, Yunhai
Zhou, Junwei
Cao, Yumo
Long, Yitao
Xu, Yiwei
Jiang, Qiyi
Wang, Weiyao
Cao, Xiaoyu
Sun, Zhen
Zou, Yiran
Du, Nan
author_facet Hu, Yunhai
Zhou, Junwei
Cao, Yumo
Long, Yitao
Xu, Yiwei
Jiang, Qiyi
Wang, Weiyao
Cao, Xiaoyu
Sun, Zhen
Zou, Yiran
Du, Nan
contents Effective retrieval in complex domains requires bridging the gap between structured metadata and unstructured content. Existing systems typically isolate these capabilities, relying on either symbolic filtering or vector similarity, failing to capture their interplay. In this work, we propose DSL-R1, a unified framework that synergizes logical reasoning with semantic matching via a novel Domain-Specific Language (DSL). By embedding vector primitives within SQL-style operators, our approach leverages the complementary strengths of symbolic precision and semantic coverage. We further introduce a reinforcement learning mechanism where rule-based execution feedback and retrieval quality rewards jointly optimize the DSL generation, balancing structural correctness and semantic alignment. Evaluations on a large-scale industrial email benchmark demonstrate that DSL-R1 achieves a +12.3% improvement in Hit@1/3, consistently outperforming decoupled baselines and establishing a robust paradigm for hybrid retrieval.
format Preprint
id arxiv_https___arxiv_org_abs_2603_21018
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle DSL-R1: From SQL to DSL for Training Retrieval Agents across Structured and Unstructured Data with Reinforcement Learning
Hu, Yunhai
Zhou, Junwei
Cao, Yumo
Long, Yitao
Xu, Yiwei
Jiang, Qiyi
Wang, Weiyao
Cao, Xiaoyu
Sun, Zhen
Zou, Yiran
Du, Nan
Information Retrieval
Artificial Intelligence
Databases
Machine Learning
Effective retrieval in complex domains requires bridging the gap between structured metadata and unstructured content. Existing systems typically isolate these capabilities, relying on either symbolic filtering or vector similarity, failing to capture their interplay. In this work, we propose DSL-R1, a unified framework that synergizes logical reasoning with semantic matching via a novel Domain-Specific Language (DSL). By embedding vector primitives within SQL-style operators, our approach leverages the complementary strengths of symbolic precision and semantic coverage. We further introduce a reinforcement learning mechanism where rule-based execution feedback and retrieval quality rewards jointly optimize the DSL generation, balancing structural correctness and semantic alignment. Evaluations on a large-scale industrial email benchmark demonstrate that DSL-R1 achieves a +12.3% improvement in Hit@1/3, consistently outperforming decoupled baselines and establishing a robust paradigm for hybrid retrieval.
title DSL-R1: From SQL to DSL for Training Retrieval Agents across Structured and Unstructured Data with Reinforcement Learning
topic Information Retrieval
Artificial Intelligence
Databases
Machine Learning
url https://arxiv.org/abs/2603.21018