Video-MOPD: Multi-Teacher On-Policy Distillation for Video Understanding

2026-09-08

AI Summary

为提升视频理解能力，通过多教师在线策略蒸馏和强化学习优化，结合可靠性感知信息采样方法，实现模型在多种视频理解任务上的性能提升。

Abstract

Video understanding demands a convergence of complementary capabilities across perception, temporal understanding, and complex reasoning, which are difficult to jointly optimize within a single model. We introduce Video-MOPD-8B, an open-weight model dedicated to video understanding tasks. To fundamentally enhance its capabilities, we conduct targeted reinforcement learning (RL) optimization across three core domains: video temporal grounding (VTG), general video comprehension, and video STEM reasoning. We then unify their complementary capabilities via Multi-Teacher On-Policy Distillation (MOPD), which consolidates expert knowledge by supervising student-generated trajectories with routed teacher feedback. We further introduce Reliability-Aware Informative Sampling (RAIS), which selects examples with consistently reliable teacher supervision and large teacher-student performance gaps. Together, these components enable Video-MOPD-8B to achieve coordinated and comprehensive performance gains across diverse video understanding tasks. Extensive experiments on comprehensive benchmarks covering general video understanding, temporal grounding, video reasoning, and video STEM tasks demonstrate that Video-MOPD-8B achieves state-of-the-art performance among existing models at a comparable scale. The trained model weights are available at https://huggingface.co/LandH/Video-MOPD-8B.

Problem

• video understanding
• perception
• temporal understanding
• complex reasoning

Innovation

• Multi-Teacher On-Policy Distillation
• Reliability-Aware Informative Sampling
• Reinforcement Learning