@inproceedings{zhang-etal-2026-deltamem,
title = "{D}elta{M}em: Towards Agentic Memory Management via Reinforcement Learning",
author = "Zhang, Qi and
Huang, Shen and
Liuchu and
Yang, Shouqing and
Hu, Xiaomeng and
Xie, Pengjun and
Wang, Haobo and
Zhao, Junbo",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1018/",
pages = "23390--23417",
ISBN = "979-8-89176-459-0",
abstract = "Recent advances in persona-centric memory demonstrate the promise of multi-agent systems for managing persona memory, especially in conversational scenarios. However, their complex pipelines often incur information loss and poor cross-scenario robustness, leading to suboptimal performance. In this paper, we propose DeltaMem, an agentic memory management system that formulates persona-centric memory management as an end-to-end task within a single-agent setting. To further improve the performance of our agentic memory manager, we draw inspiration from the evolution of human memory and synthesize a user-assistant dialogue dataset along with corresponding operation-level memory updating labels. Building on this, we introduce a novel Memory-based Levenshtein Distance to formalize the memory updating reward, and propose a tailored reinforcement learning framework to further enhance the management capabilities of DeltaMem. Extensive experiments show that both training-free and RL-trained DeltaMem outperform all product-level baselines across diverse long-term memory benchmarks, including LoCoMo, HaluMem, and PersonaMem."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2026-deltamem">
<titleInfo>
<title>DeltaMem: Towards Agentic Memory Management via Reinforcement Learning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Qi</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shen</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name>
<namePart>Liuchu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shouqing</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaomeng</namePart>
<namePart type="family">Hu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pengjun</namePart>
<namePart type="family">Xie</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haobo</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Junbo</namePart>
<namePart type="family">Zhao</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>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</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">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Recent advances in persona-centric memory demonstrate the promise of multi-agent systems for managing persona memory, especially in conversational scenarios. However, their complex pipelines often incur information loss and poor cross-scenario robustness, leading to suboptimal performance. In this paper, we propose DeltaMem, an agentic memory management system that formulates persona-centric memory management as an end-to-end task within a single-agent setting. To further improve the performance of our agentic memory manager, we draw inspiration from the evolution of human memory and synthesize a user-assistant dialogue dataset along with corresponding operation-level memory updating labels. Building on this, we introduce a novel Memory-based Levenshtein Distance to formalize the memory updating reward, and propose a tailored reinforcement learning framework to further enhance the management capabilities of DeltaMem. Extensive experiments show that both training-free and RL-trained DeltaMem outperform all product-level baselines across diverse long-term memory benchmarks, including LoCoMo, HaluMem, and PersonaMem.</abstract>
<identifier type="citekey">zhang-etal-2026-deltamem</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1018/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>23390</start>
<end>23417</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T DeltaMem: Towards Agentic Memory Management via Reinforcement Learning
%A Zhang, Qi
%A Huang, Shen
%A Yang, Shouqing
%A Hu, Xiaomeng
%A Xie, Pengjun
%A Wang, Haobo
%A Zhao, Junbo
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%A Liuchu
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F zhang-etal-2026-deltamem
%X Recent advances in persona-centric memory demonstrate the promise of multi-agent systems for managing persona memory, especially in conversational scenarios. However, their complex pipelines often incur information loss and poor cross-scenario robustness, leading to suboptimal performance. In this paper, we propose DeltaMem, an agentic memory management system that formulates persona-centric memory management as an end-to-end task within a single-agent setting. To further improve the performance of our agentic memory manager, we draw inspiration from the evolution of human memory and synthesize a user-assistant dialogue dataset along with corresponding operation-level memory updating labels. Building on this, we introduce a novel Memory-based Levenshtein Distance to formalize the memory updating reward, and propose a tailored reinforcement learning framework to further enhance the management capabilities of DeltaMem. Extensive experiments show that both training-free and RL-trained DeltaMem outperform all product-level baselines across diverse long-term memory benchmarks, including LoCoMo, HaluMem, and PersonaMem.
%U https://aclanthology.org/2026.emnlp-main.1018/
%P 23390-23417
Markdown (Informal)
[DeltaMem: Towards Agentic Memory Management via Reinforcement Learning](https://aclanthology.org/2026.emnlp-main.1018/) (Zhang et al., EMNLP 2026)
ACL
- Qi Zhang, Shen Huang, Liuchu, Shouqing Yang, Xiaomeng Hu, Pengjun Xie, Haobo Wang, and Junbo Zhao. 2026. DeltaMem: Towards Agentic Memory Management via Reinforcement Learning. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 23390–23417, Budapest, Hungary. Association for Computational Linguistics.