@inproceedings{zhou-etal-2026-validity,
title = "Validity Is Not Visibility: A Lifecycle {FSM} for Agent Memory",
author = "Zhou, Shuo and
Xu, Pengcheng and
qi, Shaotong and
Que, Li and
Sun, Ming and
Li, Han",
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.260/",
pages = "5895--5913",
ISBN = "979-8-89176-459-0",
abstract = "Long-term memory for LLM agents must answer two questions for every stored fact: whether the fact is still true and whether it should be surfaced for a given query. We refer to these two dimensions as validity and visibility, respectively. Existing memory systems often collapse these two decisions into a single scalar score. However, current and historical queries may require opposite retrieval decisions for the same fact: a superseded fact should be hidden for a current-intent query but surfaced for a historical-intent one. This creates a structural failure mode: superseded and expired facts are collapsed onto the same validity axis. To address this problem, we propose MemFSM, a training-free retrieval-side layer that separates validity from visibility through a memory-lifecycle finite-state machine. Along the validity axis, a four-state FSM tracks whether each fact has been superseded and expired. Along the visibility axis, an intent-aware router decides whether to surface or hide a fact based on its state and the query intent. By decoupling visibility from validity, MemFSM improves performance across heterogeneous memory benchmarks and diverse LLM backbones. To evaluate memory systems under fact conflicts and diverse query intents, we further introduce ValidMem, a benchmark that annotates each fact with both supersedure and expiry labels and pairs each memory store with both current-intent and historical-intent queries."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhou-etal-2026-validity">
<titleInfo>
<title>Validity Is Not Visibility: A Lifecycle FSM for Agent Memory</title>
</titleInfo>
<name type="personal">
<namePart type="given">Shuo</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pengcheng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shaotong</namePart>
<namePart type="family">qi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Li</namePart>
<namePart type="family">Que</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ming</namePart>
<namePart type="family">Sun</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Han</namePart>
<namePart type="family">Li</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>Long-term memory for LLM agents must answer two questions for every stored fact: whether the fact is still true and whether it should be surfaced for a given query. We refer to these two dimensions as validity and visibility, respectively. Existing memory systems often collapse these two decisions into a single scalar score. However, current and historical queries may require opposite retrieval decisions for the same fact: a superseded fact should be hidden for a current-intent query but surfaced for a historical-intent one. This creates a structural failure mode: superseded and expired facts are collapsed onto the same validity axis. To address this problem, we propose MemFSM, a training-free retrieval-side layer that separates validity from visibility through a memory-lifecycle finite-state machine. Along the validity axis, a four-state FSM tracks whether each fact has been superseded and expired. Along the visibility axis, an intent-aware router decides whether to surface or hide a fact based on its state and the query intent. By decoupling visibility from validity, MemFSM improves performance across heterogeneous memory benchmarks and diverse LLM backbones. To evaluate memory systems under fact conflicts and diverse query intents, we further introduce ValidMem, a benchmark that annotates each fact with both supersedure and expiry labels and pairs each memory store with both current-intent and historical-intent queries.</abstract>
<identifier type="citekey">zhou-etal-2026-validity</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.260/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>5895</start>
<end>5913</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Validity Is Not Visibility: A Lifecycle FSM for Agent Memory
%A Zhou, Shuo
%A Xu, Pengcheng
%A qi, Shaotong
%A Que, Li
%A Sun, Ming
%A Li, Han
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%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 zhou-etal-2026-validity
%X Long-term memory for LLM agents must answer two questions for every stored fact: whether the fact is still true and whether it should be surfaced for a given query. We refer to these two dimensions as validity and visibility, respectively. Existing memory systems often collapse these two decisions into a single scalar score. However, current and historical queries may require opposite retrieval decisions for the same fact: a superseded fact should be hidden for a current-intent query but surfaced for a historical-intent one. This creates a structural failure mode: superseded and expired facts are collapsed onto the same validity axis. To address this problem, we propose MemFSM, a training-free retrieval-side layer that separates validity from visibility through a memory-lifecycle finite-state machine. Along the validity axis, a four-state FSM tracks whether each fact has been superseded and expired. Along the visibility axis, an intent-aware router decides whether to surface or hide a fact based on its state and the query intent. By decoupling visibility from validity, MemFSM improves performance across heterogeneous memory benchmarks and diverse LLM backbones. To evaluate memory systems under fact conflicts and diverse query intents, we further introduce ValidMem, a benchmark that annotates each fact with both supersedure and expiry labels and pairs each memory store with both current-intent and historical-intent queries.
%U https://aclanthology.org/2026.emnlp-main.260/
%P 5895-5913
Markdown (Informal)
[Validity Is Not Visibility: A Lifecycle FSM for Agent Memory](https://aclanthology.org/2026.emnlp-main.260/) (Zhou et al., EMNLP 2026)
ACL
- Shuo Zhou, Pengcheng Xu, Shaotong qi, Li Que, Ming Sun, and Han Li. 2026. Validity Is Not Visibility: A Lifecycle FSM for Agent Memory. In Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing, pages 5895–5913, Budapest, Hungary. Association for Computational Linguistics.