VideoSEAL: Mitigating Evidence Misalignment in Agentic Long Video Understanding by Decoupling Answer Authority
For researchers in long video understanding, this work addresses a critical failure mode of existing agentic systems by decoupling answer authority from planning, enabling more reliable and interpretable reasoning.
The paper identifies 'evidence misalignment' in long video understanding agents, where correct answers lack supporting evidence, and proposes a decoupled planner-inspector framework that separates planning from answer authority. The method achieves 55.1% on LVBench and 62.0% on LongVideoBench, improving both accuracy and evidence alignment.
Long video question answering requires locating sparse, time-scattered visual evidence within highly redundant content. Although current MLLMs perform well on short videos, long videos introduce long-horizon search and verification, which often necessitates multi-turn, agentic interaction. We show that existing LVU agents can exhibit "evidence misalignment": they produce correct answers that are not supported by the retrieved or inspected evidence. To characterize this failure, we introduce two diagnostics (temporal groundedness and semantic groundedness) and use them to reveal two pressures that amplify misalignment: prompt pressure from shared-context saturation at inference time and reward pressure from outcome-only optimization during training. These findings point to a structural root cause: the coupled agent paradigm conflates long-horizon planning with answer authority. We therefore propose the decoupled planner-inspector framework, which separates planning from answer authority and gates final answering on pixel-level verification. Across four long-video benchmarks, our framework improves both answer accuracy and evidence alignment, achieving 55.1% on LVBench and 62.0% on LongVideoBench while producing interpretable search trajectories. Moreover, the decoupled architecture scales consistently with increased search budgets and supports plug-and-play upgrades of the MLLM backbone without retraining the planner. Code and models are available at https://github.com/Echochef/VideoSEAL.