Metraipolis › archive › dc_1768dc69cc52

ProRetrieval: Learning to Orchestrate Hybrid Search via Executable Program Synthesis — arXiv:2608.27017

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.27017
License (as archived)
CC0-1.0
Archived
2026-10-17T17:26:35.168Z
SHA-256 of the archived text
925308404710822e962aa02ec56bda94a140cb1b63f4205baf5bd2a2130760fd
Stable URL of this page
https://metraipolis.com/documento/dc_1768dc69cc52

How to cite

ProRetrieval: Learning to Orchestrate Hybrid Search via Executable Program Synthesis — arXiv:2608.27017 — arXiv (en), https://arxiv.org/abs/2608.27017 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1768dc69cc52 (Metraipolis archive dc_1768dc69cc52, sha256 925308404710822e962aa02ec56bda94a140cb1b63f4205baf5bd2a2130760fd)
BibTeX
@misc{dc1768dc69cc52,
  title = {ProRetrieval: Learning to Orchestrate Hybrid Search via Executable Program Synthesis — arXiv:2608.27017},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.27017},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1768dc69cc52; sha256 925308404710822e962aa02ec56bda94a140cb1b63f4205baf5bd2a2130760fd},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27017. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Real-world retrieval often composes structured constraints with semantic intents over text and images through arbitrary Boolean logic. Existing hybrid pipelines such as reciprocal rank fusion or self-querying retrievers admit only a fixed form of composition, while recent reinforcement-learning retrievers train the language model as a query generator for a single backend, leaving the orchestration of heterogeneous retrieval paths outside its action space. We propose ProRetrieval, which recasts the language model as a retrieval orchestrator: given a natural-language query, it synthesizes an executable program in a hybrid DSL interleaving SQL operators over structured fields with vector-retrieval primitives over text and images, with SQL itself providing the logical algebra that fuses heterogeneous candidate sets. We train Qwen3-4B with GRPO and DAPO under a hierarchical four-term reward, and evaluate on two new benchmarks built from Amazon products and Enron email. Our 4B model surpasses GPT-5.5 (Hit@1 0.81 vs. 0.69 on e-commerce; 0.91 vs. 0.86 on email) and Claude Opus 4.7 and a comprehensive suite of retrieval, LLM-augmented, structured-query, and graph-based baselines. Code: this https URL data: this https URL .

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-28 in cs.IR).
Categorie arXiv: Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2608.27017 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T17:18:28.088Z dalla pagina https://arxiv.org/catchup/cs.IR/2026-08-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).