RankVideo : Reasoning Reranking for Text-to-Video Retrieval

Tyler Skow, Alexander Martin, Benjamin Van Durme, Rama Chellappa, Reno Kriz · 2026

Read RankVideo : Reasoning Reranking for Text-to-Video Retrieval by Tyler Skow, Alexander Martin, Benjamin Van Durme, Rama Chellappa, Reno Kriz (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-2602-02444.txt.

Description

Tyler Skow and 4 co-authors bring step-by-step thinking to the reranking stage of text-to-video search, in a 2026 arXiv paper. Retrieval systems pair a fast first-stage retriever with a stronger model that refines the shortlist. Thinking-based reranking changed text search, but video reranking with explicit thinking was largely unexplored.

Their reranker, RANKVIDEO, examines each query-video pair using the video content itself to judge relevance. Training follows a two-stage plan: first supervised fine-tuning grounded in what the video shows, then reranking training that mixes pointwise, pairwise, and teacher-confidence objectives. A data synthesis workflow builds training pairs that demand real thought. On the large MultiVENT 2.0 benchmark, the system improved a standard ranking metric by about thirty-one percent on average. It beat text-only and vision-language alternatives while also running more efficiently.

Catalog record

CREATORTyler Skow, Alexander Martin, Benjamin Van Durme, Rama Chellappa, Reno Kriz
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-02444

Browse the full library · About the Tower of Babel