@inproceedings{zeng-etal-2026-grap,
title = "{G}ra{P}-Mem: Granularity-Aware Planning for Adaptive Memory Access in Long-Horizon {LLM} Agents",
author = "Zeng, Qinyang and
Dong, Jinhao and
LU, Wei and
Du, Xiaoyong",
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.2375/",
pages = "51016--51032",
ISBN = "979-8-89176-462-0",
abstract = "LLM agents increasingly operate in long-horizon tasks and extended dialogues, where they must retrieve useful past experiences beyond limited context windows. A central challenge is to balance efficient recall with faithful evidence recovery. Existing memory methods often rely on static access strategies: compact memories reduce context cost but may omit fine-grained evidence, whereas full histories preserve details but introduce redundant and noisy context. We propose GraP-Mem, a multi-granularity memory planning framework that adaptively schedules memory access for each question. GraP-Mem represents each historical session with two complementary views: compact semantic memories for efficient retrieval and source contexts for detailed evidence recovery. During inference, a Plan Agent generates retrieval-oriented information needs and queries, while an Integration Agent maintains a question-specific evidence state that consolidates useful evidence and identifies remaining gaps. When the evidence is incomplete, GraP-Mem selectively expands access to source contexts or performs additional retrieval. The final Answer Model is conditioned only on the integrated evidence state. Experiments on LoCoMo and NarrativeQA show that GraP-Mem consistently improves F1 and BLEU across backbone models, demonstrating the effectiveness of plan-guided memory access across multiple granularities."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zeng-etal-2026-grap">
<titleInfo>
<title>GraP-Mem: Granularity-Aware Planning for Adaptive Memory Access in Long-Horizon LLM Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Qinyang</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jinhao</namePart>
<namePart type="family">Dong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Wei</namePart>
<namePart type="family">LU</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaoyong</namePart>
<namePart type="family">Du</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>LLM agents increasingly operate in long-horizon tasks and extended dialogues, where they must retrieve useful past experiences beyond limited context windows. A central challenge is to balance efficient recall with faithful evidence recovery. Existing memory methods often rely on static access strategies: compact memories reduce context cost but may omit fine-grained evidence, whereas full histories preserve details but introduce redundant and noisy context. We propose GraP-Mem, a multi-granularity memory planning framework that adaptively schedules memory access for each question. GraP-Mem represents each historical session with two complementary views: compact semantic memories for efficient retrieval and source contexts for detailed evidence recovery. During inference, a Plan Agent generates retrieval-oriented information needs and queries, while an Integration Agent maintains a question-specific evidence state that consolidates useful evidence and identifies remaining gaps. When the evidence is incomplete, GraP-Mem selectively expands access to source contexts or performs additional retrieval. The final Answer Model is conditioned only on the integrated evidence state. Experiments on LoCoMo and NarrativeQA show that GraP-Mem consistently improves F1 and BLEU across backbone models, demonstrating the effectiveness of plan-guided memory access across multiple granularities.</abstract>
<identifier type="citekey">zeng-etal-2026-grap</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2375/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>51016</start>
<end>51032</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T GraP-Mem: Granularity-Aware Planning for Adaptive Memory Access in Long-Horizon LLM Agents
%A Zeng, Qinyang
%A Dong, Jinhao
%A LU, Wei
%A Du, Xiaoyong
%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 zeng-etal-2026-grap
%X LLM agents increasingly operate in long-horizon tasks and extended dialogues, where they must retrieve useful past experiences beyond limited context windows. A central challenge is to balance efficient recall with faithful evidence recovery. Existing memory methods often rely on static access strategies: compact memories reduce context cost but may omit fine-grained evidence, whereas full histories preserve details but introduce redundant and noisy context. We propose GraP-Mem, a multi-granularity memory planning framework that adaptively schedules memory access for each question. GraP-Mem represents each historical session with two complementary views: compact semantic memories for efficient retrieval and source contexts for detailed evidence recovery. During inference, a Plan Agent generates retrieval-oriented information needs and queries, while an Integration Agent maintains a question-specific evidence state that consolidates useful evidence and identifies remaining gaps. When the evidence is incomplete, GraP-Mem selectively expands access to source contexts or performs additional retrieval. The final Answer Model is conditioned only on the integrated evidence state. Experiments on LoCoMo and NarrativeQA show that GraP-Mem consistently improves F1 and BLEU across backbone models, demonstrating the effectiveness of plan-guided memory access across multiple granularities.
%U https://aclanthology.org/2026.findings-emnlp.2375/
%P 51016-51032
Markdown (Informal)
[GraP-Mem: Granularity-Aware Planning for Adaptive Memory Access in Long-Horizon LLM Agents](https://aclanthology.org/2026.findings-emnlp.2375/) (Zeng et al., Findings 2026)
ACL