@inproceedings{abaskohi-etal-2026-sproutrag,
title = "{S}prout{RAG}: Attention-Guided Tree Search with Progressive Embeddings for Long-Document {RAG}",
author = "Abaskohi, Amirhossein and
Laradji, Issam H. and
West, Peter and
Carenini, Giuseppe",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2175/",
pages = "46521--46535",
ISBN = "979-8-89176-462-0",
abstract = "Retrieval-augmented generation (RAG) systems must balance retrieval granularity with contextual coherence, a challenge that existing methods address through LLM-guided chunking, single-level context expansion, or hierarchical summarization. These approaches variously depend on costly LLM calls during indexing or retrieval, limit context aggregation to a single granularity level, or introduce information loss through summarization. We present SproutRAG, an attention-guided hierarchical RAG framework that builds a binary tree over sentence-level chunks using learned inter-sentence attention from a sentence-level large language model (SLLM). To avoid proximity bias from uniform attention averaging, SproutRAG learns head- and layer-level aggregation weights that identify attention patterns most aligned with semantic document structure. At retrieval time, it uses hierarchical beam search over the tree to gather candidates at multiple granularities, capturing multi-sentence relevance beyond flat retrieval. The framework is trained end-to-end with a joint objective that improves both embeddings and tree structure, without requiring external LLM calls. Experiments across four benchmarks spanning scientific, legal, and open-domain settings demonstrate that SproutRAG improves information efficiency (IE) by 6.1{\%} on average over the strongest baseline."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="abaskohi-etal-2026-sproutrag">
<titleInfo>
<title>SproutRAG: Attention-Guided Tree Search with Progressive Embeddings for Long-Document RAG</title>
</titleInfo>
<name type="personal">
<namePart type="given">Amirhossein</namePart>
<namePart type="family">Abaskohi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Issam</namePart>
<namePart type="given">H</namePart>
<namePart type="family">Laradji</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">West</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Giuseppe</namePart>
<namePart type="family">Carenini</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Retrieval-augmented generation (RAG) systems must balance retrieval granularity with contextual coherence, a challenge that existing methods address through LLM-guided chunking, single-level context expansion, or hierarchical summarization. These approaches variously depend on costly LLM calls during indexing or retrieval, limit context aggregation to a single granularity level, or introduce information loss through summarization. We present SproutRAG, an attention-guided hierarchical RAG framework that builds a binary tree over sentence-level chunks using learned inter-sentence attention from a sentence-level large language model (SLLM). To avoid proximity bias from uniform attention averaging, SproutRAG learns head- and layer-level aggregation weights that identify attention patterns most aligned with semantic document structure. At retrieval time, it uses hierarchical beam search over the tree to gather candidates at multiple granularities, capturing multi-sentence relevance beyond flat retrieval. The framework is trained end-to-end with a joint objective that improves both embeddings and tree structure, without requiring external LLM calls. Experiments across four benchmarks spanning scientific, legal, and open-domain settings demonstrate that SproutRAG improves information efficiency (IE) by 6.1% on average over the strongest baseline.</abstract>
<identifier type="citekey">abaskohi-etal-2026-sproutrag</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2175/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>46521</start>
<end>46535</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SproutRAG: Attention-Guided Tree Search with Progressive Embeddings for Long-Document RAG
%A Abaskohi, Amirhossein
%A Laradji, Issam H.
%A West, Peter
%A Carenini, Giuseppe
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F abaskohi-etal-2026-sproutrag
%X Retrieval-augmented generation (RAG) systems must balance retrieval granularity with contextual coherence, a challenge that existing methods address through LLM-guided chunking, single-level context expansion, or hierarchical summarization. These approaches variously depend on costly LLM calls during indexing or retrieval, limit context aggregation to a single granularity level, or introduce information loss through summarization. We present SproutRAG, an attention-guided hierarchical RAG framework that builds a binary tree over sentence-level chunks using learned inter-sentence attention from a sentence-level large language model (SLLM). To avoid proximity bias from uniform attention averaging, SproutRAG learns head- and layer-level aggregation weights that identify attention patterns most aligned with semantic document structure. At retrieval time, it uses hierarchical beam search over the tree to gather candidates at multiple granularities, capturing multi-sentence relevance beyond flat retrieval. The framework is trained end-to-end with a joint objective that improves both embeddings and tree structure, without requiring external LLM calls. Experiments across four benchmarks spanning scientific, legal, and open-domain settings demonstrate that SproutRAG improves information efficiency (IE) by 6.1% on average over the strongest baseline.
%U https://aclanthology.org/2026.findings-emnlp.2175/
%P 46521-46535
Markdown (Informal)
[SproutRAG: Attention-Guided Tree Search with Progressive Embeddings for Long-Document RAG](https://aclanthology.org/2026.findings-emnlp.2175/) (Abaskohi et al., Findings 2026)
ACL