aiqbench analysis
This research addresses the critical gap in current Multimodal Large Language Models (MLLMs) by introducing a new benchmark and dataset for One-to-Many Temporal Grounding, a task essential for handling multiple disjoint video segments per query. The authors propose novel reward functions leveraging Chain-of-Thought reasoning to improve event cardinality perception, achieving a significant performance lift over leading commercial models. For practitioners, this work establishes new evaluation metrics like Count Accuracy and Effective Temporal F1 that are necessary for accurately assessing progress in complex video understanding scenarios.
Towards One-to-Many Temporal Grounding
arXiv:2606.06294v1 Announce Type: cross Abstract: Temporal Grounding (TG) aims to localize video segments corresponding to a textual query. Prior research predominantly focuses on single-segment retrieval. Real-world scenarios, however, often require localizing multiple disjoint segments for a single query -- a setting we term One-to-Many Temporal Grounding (OMTG). Previous state-of-the-art MLLMs, optimized for one-to-one settings, struggle in this context, often yielding near-zero scores due to a lack of event cardinality perception. To bridge this gap, we present a systematic solution with three key contributions. First, we establish the first comprehensive OMTG benchmark, introducing Count Accuracy (C-Acc) and Effective Temporal F1 (EtF1) as evaluation metrics. Second, we curate a high-quality OMTG dataset comprising 56k samples through a sophisticated construction pipeline. Third, we develop novel temporal and caption reward functions specifically designed for OMTG. In particular, the caption reward leverages Chain-of-Thought reasoning over dense video captions to explicitly guide policy optimization toward both preciseness and completeness. Extensive experiments show our model achieves a new state-of-the-art EtF1 of 43.65\% on OMTG Bench, outperforming Gemini 2.5 Pro and Seed-1.8 by 15.85\% and 15.61\%, respectively.
Abstract:Temporal Grounding (TG) aims to localize video segments corresponding to a textual query. Prior research predominantly focuses on single-segment retrieval. Real-world scenarios, however, often require localizing multiple disjoint segments for a single query -- a setting we term One-to-Many Temporal Grounding (OMTG). Previous state-of-the-art MLLMs, optimized for one-to-one settings, struggle in this context, often yielding near-zero scores due to a lack of event cardinality perception. To bridge this gap, we present a systematic solution with three key contributions. First, we establish the first comprehensive OMTG benchmark, introducing Count Accuracy (C-Acc) and Effective Temporal F1 (EtF1) as evaluation metrics. Second, we curate a high-quality OMTG dataset comprising 56k samples through a sophisticated construction pipeline. Third, we develop novel temporal and caption reward functions specifically designed for OMTG. In particular, the caption reward leverages Chain-of-Thought reasoning over dense video captions to explicitly guide policy optimization toward both preciseness and completeness. Extensive experiments show our model achieves a new state-of-the-art EtF1 of 43.65\% on OMTG Bench, outperforming Gemini 2.5 Pro and Seed-1.8 by 15.85\% and 15.61\%, respectively.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2606.06294 [cs.CV]
(or arXiv:2606.06294v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2606.06294 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 274 words
Related coverage
Game Design Creativity evals
LessWrong · Oct 12, 2026
Multi-Agent Self-Improving Reinforcement Learning for Video Reasoning
arXiv · Sep 1, 2026
Learning and Clustering on Temporal Graphs: Principles, Primitives, and Pooling
arXiv · Aug 5, 2026
Light Multiplication Unlocks Faster Quantum Computations
Quantum Zeitgeist · Jun 16, 2026