@inproceedings{yu-etal-2026-memora,
title = "{MEMORA}: Embodied Action Memory from Egocentric Videos for Reasoning and Planning",
author = "Yu, Zihao and
Yuan, Xiu and
Zhang, Chongjie",
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.1211/",
doi = "10.18653/2026.findings-emnlp.1211",
pages = "25551--25600",
ISBN = "979-8-89176-462-0",
abstract = "Embodied agents must reason and act from experience accumulated before future questions or goals are known, yet existing memory systems often treat such experience as flat text, retrieved clips, or static graphs. We introduce \textbf{MEMORA}, an embodied memory framework that forms persistent participant-specific world memory from continuous egocentric video. MEMORA instantiates a formation{--}consolidation lifecycle: \textit{online memory editing} maintains typed stores for spatial context, object identity and state, and temporal action evidence, while \textit{offline consolidation} abstracts repeated experience into reusable routines, habits, and preferences. We evaluate MEMORA with \textbf{MEMORA-Bench}, a 45-hour egocentric-video suite for retrospective memory faithfulness and prospective memory-grounded planning. Across four open-weight answer models, MEMORA improves over no-memory, flat-text, and graph-structured memory baselines under matched perception and retrieval. The gains are largest on out-of-distribution planning, where MEMORA improves robot-grounded planning score by up to , showing that formed memory can compose new goals from remembered procedures, objects, and preferences. A physical-robot demonstration further suggests that memory formed from human egocentric video can guide participant-grounded action without in-domain robot data."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yu-etal-2026-memora">
<titleInfo>
<title>MEMORA: Embodied Action Memory from Egocentric Videos for Reasoning and Planning</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zihao</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiu</namePart>
<namePart type="family">Yuan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chongjie</namePart>
<namePart type="family">Zhang</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>Embodied agents must reason and act from experience accumulated before future questions or goals are known, yet existing memory systems often treat such experience as flat text, retrieved clips, or static graphs. We introduce MEMORA, an embodied memory framework that forms persistent participant-specific world memory from continuous egocentric video. MEMORA instantiates a formation–consolidation lifecycle: online memory editing maintains typed stores for spatial context, object identity and state, and temporal action evidence, while offline consolidation abstracts repeated experience into reusable routines, habits, and preferences. We evaluate MEMORA with MEMORA-Bench, a 45-hour egocentric-video suite for retrospective memory faithfulness and prospective memory-grounded planning. Across four open-weight answer models, MEMORA improves over no-memory, flat-text, and graph-structured memory baselines under matched perception and retrieval. The gains are largest on out-of-distribution planning, where MEMORA improves robot-grounded planning score by up to , showing that formed memory can compose new goals from remembered procedures, objects, and preferences. A physical-robot demonstration further suggests that memory formed from human egocentric video can guide participant-grounded action without in-domain robot data.</abstract>
<identifier type="citekey">yu-etal-2026-memora</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1211</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1211/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>25551</start>
<end>25600</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MEMORA: Embodied Action Memory from Egocentric Videos for Reasoning and Planning
%A Yu, Zihao
%A Yuan, Xiu
%A Zhang, Chongjie
%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 yu-etal-2026-memora
%X Embodied agents must reason and act from experience accumulated before future questions or goals are known, yet existing memory systems often treat such experience as flat text, retrieved clips, or static graphs. We introduce MEMORA, an embodied memory framework that forms persistent participant-specific world memory from continuous egocentric video. MEMORA instantiates a formation–consolidation lifecycle: online memory editing maintains typed stores for spatial context, object identity and state, and temporal action evidence, while offline consolidation abstracts repeated experience into reusable routines, habits, and preferences. We evaluate MEMORA with MEMORA-Bench, a 45-hour egocentric-video suite for retrospective memory faithfulness and prospective memory-grounded planning. Across four open-weight answer models, MEMORA improves over no-memory, flat-text, and graph-structured memory baselines under matched perception and retrieval. The gains are largest on out-of-distribution planning, where MEMORA improves robot-grounded planning score by up to , showing that formed memory can compose new goals from remembered procedures, objects, and preferences. A physical-robot demonstration further suggests that memory formed from human egocentric video can guide participant-grounded action without in-domain robot data.
%R 10.18653/2026.findings-emnlp.1211
%U https://aclanthology.org/2026.findings-emnlp.1211/
%U https://doi.org/10.18653/2026.findings-emnlp.1211
%P 25551-25600
Markdown (Informal)
[MEMORA: Embodied Action Memory from Egocentric Videos for Reasoning and Planning](https://aclanthology.org/2026.findings-emnlp.1211/) (Yu et al., Findings 2026)
ACL