@inproceedings{mahmoud-etal-2026-memora,
title = "{MEMORA}: A Memory-Enhanced Multimodal Committee Reranking Agent for Reasoning-Intensive Retrieval",
author = "Mahmoud, Mohamed and
Senussi, Mostafa Farouk and
Abdalla, Mahmoud and
Kasem, Mahmoud SalahEldin and
Abdallah, Abdelrahman and
Kang, Hyun Soo",
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.46/",
pages = "1017--1028",
ISBN = "979-8-89176-459-0",
abstract = "Reasoning-intensive retrieval is fundamentally challenged when queries demand deep semantic inference and arrive in multimodal form. Existing systems treat each query in isolation, rely on monolithic rerankers, and lack retrievers trained for cross-modal reasoning. We introduce \textbf{MEMORA}, a \textbf{M}emory-\textbf{E}nhanced \textbf{M}ultimodal \textbf{O}rchestrated \textbf{R}eranking \textbf{A}gent built around three contributions: (1)\textbf{Collaborative Factor-Committee Reranking}, which decomposes queries into intent factors, extracts factor-specific evidence into a structured matrix, and synthesizes judgments via a memory-conditioned meta-verdict model; (2) an \textbf{Episodic Memory Module} that tracks factor outcomes across queries and dynamically adjusts priorities at inference time; and (3) a \textbf{Reasoning-Aware Dense Retriever} fine-tuned with chain-of-thought augmentation on 20k$+$ reasoning-intensive pairs. Evaluated on MM-BRIGHT across 29 domains, \textsc{MEMORA} achieves nDCG@10 of \textbf{41.5} (text-to-text) and \textbf{38.9} (multimodal query-to-text), outperforming the strongest agentic text-only baseline by \textbf{+8.3} points and the strongest agentic multimodal baseline by \textbf{+7.9} points."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mahmoud-etal-2026-memora">
<titleInfo>
<title>MEMORA: A Memory-Enhanced Multimodal Committee Reranking Agent for Reasoning-Intensive Retrieval</title>
</titleInfo>
<name type="personal">
<namePart type="given">Mohamed</namePart>
<namePart type="family">Mahmoud</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mostafa</namePart>
<namePart type="given">Farouk</namePart>
<namePart type="family">Senussi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mahmoud</namePart>
<namePart type="family">Abdalla</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mahmoud</namePart>
<namePart type="given">SalahEldin</namePart>
<namePart type="family">Kasem</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Abdelrahman</namePart>
<namePart type="family">Abdallah</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hyun</namePart>
<namePart type="given">Soo</namePart>
<namePart type="family">Kang</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>Reasoning-intensive retrieval is fundamentally challenged when queries demand deep semantic inference and arrive in multimodal form. Existing systems treat each query in isolation, rely on monolithic rerankers, and lack retrievers trained for cross-modal reasoning. We introduce MEMORA, a Memory-Enhanced Multimodal Orchestrated Reranking Agent built around three contributions: (1)Collaborative Factor-Committee Reranking, which decomposes queries into intent factors, extracts factor-specific evidence into a structured matrix, and synthesizes judgments via a memory-conditioned meta-verdict model; (2) an Episodic Memory Module that tracks factor outcomes across queries and dynamically adjusts priorities at inference time; and (3) a Reasoning-Aware Dense Retriever fine-tuned with chain-of-thought augmentation on 20k+ reasoning-intensive pairs. Evaluated on MM-BRIGHT across 29 domains, MEMORA achieves nDCG@10 of 41.5 (text-to-text) and 38.9 (multimodal query-to-text), outperforming the strongest agentic text-only baseline by +8.3 points and the strongest agentic multimodal baseline by +7.9 points.</abstract>
<identifier type="citekey">mahmoud-etal-2026-memora</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.46/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1017</start>
<end>1028</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T MEMORA: A Memory-Enhanced Multimodal Committee Reranking Agent for Reasoning-Intensive Retrieval
%A Mahmoud, Mohamed
%A Senussi, Mostafa Farouk
%A Abdalla, Mahmoud
%A Kasem, Mahmoud SalahEldin
%A Abdallah, Abdelrahman
%A Kang, Hyun Soo
%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 mahmoud-etal-2026-memora
%X Reasoning-intensive retrieval is fundamentally challenged when queries demand deep semantic inference and arrive in multimodal form. Existing systems treat each query in isolation, rely on monolithic rerankers, and lack retrievers trained for cross-modal reasoning. We introduce MEMORA, a Memory-Enhanced Multimodal Orchestrated Reranking Agent built around three contributions: (1)Collaborative Factor-Committee Reranking, which decomposes queries into intent factors, extracts factor-specific evidence into a structured matrix, and synthesizes judgments via a memory-conditioned meta-verdict model; (2) an Episodic Memory Module that tracks factor outcomes across queries and dynamically adjusts priorities at inference time; and (3) a Reasoning-Aware Dense Retriever fine-tuned with chain-of-thought augmentation on 20k+ reasoning-intensive pairs. Evaluated on MM-BRIGHT across 29 domains, MEMORA achieves nDCG@10 of 41.5 (text-to-text) and 38.9 (multimodal query-to-text), outperforming the strongest agentic text-only baseline by +8.3 points and the strongest agentic multimodal baseline by +7.9 points.
%U https://aclanthology.org/2026.emnlp-main.46/
%P 1017-1028
Markdown (Informal)
[MEMORA: A Memory-Enhanced Multimodal Committee Reranking Agent for Reasoning-Intensive Retrieval](https://aclanthology.org/2026.emnlp-main.46/) (Mahmoud et al., EMNLP 2026)
ACL