MemMA: Coordinating the Memory Cycle through Multi-Agent Reasoning and In-Situ Self-Evolution

Minhua Lin; Zhiwei Zhang; Hanqing Lu; Hui Liu; Xianfeng Tang; Qi He; Xiang Zhang; Suhang Wang

cs.AIMar 19, 2026

MemMA: Coordinating the Memory Cycle through Multi-Agent Reasoning and In-Situ Self-Evolution

Authors: Minhua Lin, Zhiwei Zhang, Hanqing Lu, Hui Liu, Xianfeng Tang, Qi He, Xiang Zhang, Suhang Wang

Organizations: The Pennsylvania State University · Amazon · Microsoft

Abstract

Memory-augmented LLM agents maintain external memory banks to support long-horizon interaction, yet most existing systems treat construction, retrieval, and utilization as isolated subroutines. This creates two coupled challenges: strategic blindness on the forward path of the memory cycle, where construction and retrieval are driven by local heuristics rather than explicit strategic reasoning, and sparse, delayed supervision on the backward path, where downstream failures rarely translate into direct repairs of the memory bank. To address these challenges, we propose MemMA, a plug-and-play multi-agent framework that coordinates the memory cycle along both the forward and backward paths. On the forward path, a Meta-Thinker produces structured guidance that steers a Memory Manager during construction and directs a Query Reasoner during iterative retrieval. On the backward path, MemMA introduces in-situ self-evolving memory construction, which synthesizes probe QA pairs, verifies the current memory, and converts failures into repair actions before the memory is finalized. Extensive experiments on LoCoMo show that MemMA consistently outperforms existing baselines across multiple LLM backbones and improves three different storage backends in a plug-and-play manner. Our code is publicly available at https://github.com/ventr1c/memma.

Figures & tables

Figure 1: Two challenges in leveraging the memory cycle effect.

Method	F1	B1	ACC
Static Baseline	22.64	17.24	52.60
Unguided Active	23.49	18.36	54.60
Strategic Active	24.78	17.73	59.21

Table 1: Preliminary analysis results (%) in LoCoMo dataset, GPT-4o-mini is the backbone LLM.

Model	Method	Multi-Hop			Temporal			Open-Domain			Single-Hop			Overall
Model	Method	F1	B1	ACC	F1	B1	ACC	F1	B1	ACC	F1	B1	ACC	F1	B1	ACC
GPT	Full Text	29.41	21.16	43.75	29.95	19.33	51.35	18.25	19.56	61.54	41.45	29.96	74.29	34.13	24.63	61.18
	Naive RAG	15.84	9.50	31.25	17.30	12.36	35.14	17.40	16.65	46.15	39.32	30.35	58.57	27.14	20.41	46.05
	LangMem	12.55	9.22	25.00	15.23	11.53	21.62	14.91	14.03	38.46	23.52	17.59	35.71	18.46	14.05	30.26
	A-Mem	15.56	10.88	31.25	55.01	42.40	51.35	18.18	15.27	53.85	42.72	32.43	62.86	37.90	28.85	52.63
	LightMem	33.74	29.33	65.62	59.76	51.12	78.38	31.85	24.23	76.92	43.88	34.68	78.57	44.58	36.66	75.66

Table 2: Results on LoCoMo across four question categories (multi-hop, temporal, open-domain, single-hop). We report F1, B1, and ACC (%). Best results are in bold . GPT-4o-mini and Claude-Haiku-4.5 are backbones; GPT-4o-mini is the answer agent.

\textnormal{{MemMA}}_{\mathrm{LM}}

uses LightMem as storage backend.

Method	F1	B1	ACC
Single-Agent	22.64	17.24	52.60
$\textnormal{{MemMA}}_{\mathrm{SA}}$	23.64	12.94	84.87
A-Mem	37.90	28.85	52.63
$\textnormal{{MemMA}}_{\mathrm{AM}}$	46.23	35.13	78.29
LightMem	44.58	36.66	75.66
$\textnormal{{MemMA}}_{\mathrm{LM}}$	49.40	38.28	81.58

Table 3: Flexibility across backends on LoCoMo under GPT-4o-mini. Best results per backend are in bold .

Panel of Figure 3 — Figure 3: Ablation studies of $\textnormal{{MemMA}}_{\mathrm{SA}}$ under GPT-4o-mini and Claude-Haiku-4.5 on LoCoMo.

Panel of Figure 4 — Figure 4: Impact of retrieval budget $k$ of MemMA under both GPT-4o-mini and Claude-Haiku-4.5.

Panel of Figure 5 — Figure 5: Impact of refinement budget $H$ of MemMA .

Method	Coverage	Hit@ $\geq$ 80%
NaiveRAG	42.2	24.0
A-Mem	51.0	26.0
LightMem	61.9	39.0
MemMA LM , $E_{0}$	65.2	45.2
MemMA LM , $E_{H}$	66.4	45.9

Table 4: Retrieval-process diagnostics (%) on LoCoMo.

E_{0}

and

E_{H}

denote the initial and final accumulated evidence, respectively.

Method	ACC
FullContext	50.0
NaiveRAG	58.0
A-Mem	72.0
LightMem	82.0
MemMA LM	94.0

Table 5: Results on the 20% subset of LongMemEval-S.

Method	Memory Scope	Pass Rate
Vanilla	None	62.0
MemMA /E	Within-task	66.0
MemMA	Within-task	72.0
MemMA /E	Within + cross-task	72.0
MemMA	Within + cross-task	76.0

Table 6: Results on SWE-bench-Verified-mini.

Appendix figures & tables12 assets

Supplementary material from the paper’s appendix.

Appendix

Panel of Table 7 — Table 7: The prompt template used for LLM-as-a-Judge evaluation.

Type	Example Question	Example Answer
Single-hop	What type of support group did I tell Melanie I attended recently?	An LGBTQ support group
Multi-hop	What is Melanie’s hobby for creative expression and relaxation, and when did she create the specific piece she showed me?	Melanie paints as her hobby for creative expression and relaxation. She painted a lake sunrise last year that she showed me.
Temporal	On what date and time did I have the conversation with Melanie about attending the LGBTQ support group and my career interests in counseling?	At 1:56 pm on May 8, 2023

Appendix

Table 8: Synthetic QA probe types used during probe generation in LoCoMo, with representative examples from the generated probe data.

Probe Model	F1	B1	ACC
Claude-Haiku-4.5	44.98	35.69	74.34
Claude-Sonnet-4.5	43.30	32.74	74.34
Claude-Opus-4.5	45.10	35.66	76.97

Appendix

Table 9: Impact of probe generation model on

\textnormal{{MemMA}}_{\mathrm{LM}}

with Claude-Haiku-4.5 as the construction backbone. Best results are in bold .

Probe Model	Avg. Q Len.	Avg. A Len.	One-word / $\leq$ 3	Single-hop	Multi-hop	Temporal
Haiku	18.48	19.44	4 / 15	55	25	15
Sonnet	15.42	11.13	11 / 33	64	16	15
Opus	17.38	21.55	4 / 8	58	26	11

Appendix

Table 10: Probe statistics across generation models. “One-word /

\leq

3” counts one-word and short (

\leq 3

words) answers out of

95

total per model. Question type counts follow the taxonomy in Appendix E.1 .

Model	Question	Answer
Haiku	What has the support group I attended done for my personal development and self-acceptance?	The support group has made me feel accepted and given me courage to embrace myself.
Sonnet	What did the LGBTQ support group help me feel that gave me courage to embrace myself?	Accepted
Opus	How has attending the LGBTQ support group influenced my personal growth and willingness to be open about my identity?	The support group has been a safe space that made me feel accepted, giving me the courage to embrace myself and be more open about my identity in other areas of life.

Appendix

Table 11: Representative probe QA pairs from the same dialogue session. Sonnet’s single-word answer tests only keyword presence, while Haiku and Opus require multi-attribute recall.

Panel of Table 12 — Table 12: The prompt template used for Meta-Thinker construction guidance.

Panel of Table 13 — Table 13: The prompt template used for Meta-Thinker answerability checking (Part 1).

Panel of Table 14 — Table 14: The prompt template used for Meta-Thinker answerability checking (Part 2).

Panel of Table 15 — Table 15: The prompt template used for Query Reasoner $\pi_{r}$ to generate orthogonal query $u_{h+1}$ .

Panel of Table 16 — Table 16: The prompt template used for evidence-grounded repair in self-evolution (Part 1).

Panel of Table 17 — Table 17: The prompt template used for evidence-grounded repair in self-evolution (Part 2: output format and examples).

Panel of Table 18 — Table 18: The prompt template used for semantic consolidation (deduplication) in self-evolution.

Agent Memory Management

Agent memory management studies how AI agents store, update, retrieve, compress, and forget information across long-running tasks.

See all 446 papers

Remember Before You're Asked: MemDream for Self-Probing Memory Evolution
Sep 28, 2026Mingfei Lu, Mengjia Wu, Runsong Jia +2LLM Agent Memory Graph-Based Agent Memory
Australian Artificial Intelligence Institute (AAII) University of Technology Sydney
Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents
Jun 4, 2026Shuo Ji, Yibo Li, Bryan HooiAgent Memory LLM Agent Memory
National University of Singapore, Singapore.
AdMem: Advanced Memory for Task-solving Agents
Jun 5, 2026Runzhe Wang, Huilin Lu, Shengjie Liu +2Continual Learning for LLM Agents LLM Agent Memory
Princeton University · Amazon · Arm

Sep 28, 2026J/K move · Enter open · S save

Remember Before You're Asked: MemDream for Self-Probing Memory Evolution

Mingfei Lu, Mengjia Wu, Runsong Jia, Zhe Luo+1

Australian Artificial Intelligence Institute (AAII) University of Technology Sydney

LLM Agent Memory Graph-Based Agent Memory

Open paper AskPDF arXiv

Memory is essential for enabling LLM-based agents to maintain coherent, personalized behavior over long-horizon interactions. However, existing memory systems share a fundamental limitation: they never proactively test their own memory, repairing it only after real queries expose weaknesses. This reactive paradigm means every retrieval failure corresponds to a real interaction in which the cost has already been paid. We propose MemDream, a framework that enables self-probing memory evolution for LLM agents. Our framework periodically enters offline dream cycles where three specialized agents (Dreamer, Analyst, Consolidator) collaboratively probe, diagnose, and repair the memory graph before failures occur. A policy trained via Group Relative Policy Optimization learns which repair operations produce durable retrieval improvements, while a soft decay mechanism provides reversible forgetting driven by the same anticipatory signal. Experiments on LoCoMo and MemoryAgentBench demonstrate that MemDream improves answer F1 by 4.5 points on LoCoMo and achieves a 9.1-point higher overall score on MAB over the strongest reactive-evolution baselines.

Abstract

Agent Memory Management

Explore similar work

Remember Before You're Asked: MemDream for Self-Probing Memory Evolution

Memory is Reconstructed, Not Retrieved: Graph Memory for LLM Agents

AdMem: Advanced Memory for Task-solving Agents