MUSEG: Reinforcing Video Temporal Understanding via Timestamp-Aware Multi-Segment Grounding

TL;DR

MUSEG is a reinforcement learning method that improves multimodal large language models' ability to perform fine-grained temporal reasoning in videos by using timestamp-aware multi-segment grounding.

Contribution

It introduces a novel RL-based approach with phased rewards for better temporal understanding and aligns queries with multiple video segments.

Findings

MUSEG significantly outperforms existing methods on temporal grounding tasks.

It generalizes well across diverse temporal understanding scenarios.

The approach enhances the reasoning capabilities of multimodal large language models.

Abstract

Video temporal understanding is crucial for multimodal large language models (MLLMs) to reason over events in videos. Despite recent advances in general video understanding, current MLLMs still struggle with fine-grained temporal reasoning. While reinforcement learning (RL) has been explored to address this issue recently, existing RL approaches remain limited in performance on time-sensitive tasks. In this work, we propose MUSEG, a novel RL-based method that enhances temporal understanding by introducing timestamp-aware multi-segment grounding. MUSEG enables MLLMs to align queries with multiple relevant video segments, promoting more comprehensive temporal reasoning. To facilitate effective learning, we design a customized RL training recipe with phased rewards that progressively guides the model toward temporally grounded reasoning. Extensive experiments on temporal grounding and time-sensitive video question answering (QA) tasks demonstrate that MUSEG significantly outperforms existing methods and generalizes well across diverse temporal understanding scenarios. View our project at https://github.com/THUNLP-MT/MUSEG.