BRIDGE: Multimodal-to-Text Retrieval via Reinforcement-Learned Query Alignment

Mohamed Darwish Mounis, Mohamed Mahmoud, Shaimaa Sedek, Mahmoud Abdalla, Mahmoud SalahEldin Kasem, Abdelrahman Abdallah, Hyun-Soo Kang, [0.5em] High institute for computer & information systems Chungbuk National University · 2026

Read BRIDGE: Multimodal-to-Text Retrieval via Reinforcement-Learned Query Alignment by Mohamed Darwish Mounis, Mohamed Mahmoud, Shaimaa Sedek, Mahmoud Abdalla, Mahmoud SalahEldin Kasem, Abdelrahman Abdallah, Hyun-Soo Kang, [0.5em] High institute for computer & information systems Chungbuk National University (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-07201.txt.

Description

Mounis and six colleagues address multimodal search, where a query mixing images and words must find plain-text documents. Their BRIDGE system first rewrites the noisy query into a compact search string with a reinforcement-trained aligner. A reasoning-tuned dense retriever then runs on that cleaned query. On a benchmark of 2803 queries across 29 domains, the full system beats vision-language encoders, and the aligner alone lifts a leading encoder past the best text-only retriever.

Catalog record

CREATORMohamed Darwish Mounis, Mohamed Mahmoud, Shaimaa Sedek, Mahmoud Abdalla, Mahmoud SalahEldin Kasem, Abdelrahman Abdallah, Hyun-Soo Kang, [0.5em] High institute for computer & information systems Chungbuk National University
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32827
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.07201

Browse the full library · About the Tower of Babel