MMMORRF: Multimodal Multilingual Modularized Reciprocal Rank Fusion

Authors:
  • SSSaron Samuel
    Saron Samuel
  • DDDan DeGenaro
    Dan DeGenaro
  • JGJimena Guallar-Blasco
    Jimena Guallar-Blasco
  • KSKate Sanders
    Kate Sanders
  • OEOluwaseun Eisape
    Oluwaseun Eisape
  • TSTanner Spendlove
    Tanner Spendlove
  • ARArun Reddy
    Arun Reddy
  • AMAlexander Martin
    Alexander Martin
  • AYAndrew Yates
    Andrew Yates
  • EYEugene Yang
    Eugene Yang
  • CCCameron Carpenter
    Cameron Carpenter
  • DEDavid Etter
    David Etter

Abstract
Videos inherently contain multiple modalities, including visual events, text overlays, sounds, and speech, all of which are important for retrieval. However, state-of-the-art multimodal language models like VAST and LanguageBind are built on vision-language models (VLMs), and thus overly prioritize visual signals. Retrieval benchmarks further reinforce this bias by focusing on visual queries and neglecting other modalities. We create a search system MMMORRF that extracts text and features from both visual and audio modalities and integrates them with a novel modality-aware weighted reciprocal rank fusion. MMMORRF is both effective and efficient, demonstrating practicality in searching videos based on users' information needs instead of visual descriptive queries. We evaluate MMMORRF on MultiVENT 2.0 and TVR, two multimodal benchmarks designed for more targeted information needs, and find that it improves nDCG@20 by 81% over leading multimodal encoders and 37% over single-modality retrieval, demonstrating the value of integrating diverse modalities.