Abstract
Large Video Language Models (LVLMs) have rapidly emerged as the focus of multimedia AI research. Nonetheless, when confronted with lengthy videos, these models struggle: their temporal windows are narrow, and they fail to notice fine-grained semantic shifts that unfold over extended durations. Moreover, mainstream text-based retrieval pipelines, which rely chiefly on surface-level lexical overlap, ignore the rich temporal interdependence among visual, audio, and subtitle channels. To mitigate these limitations, we propose TV-RAG, a training-free architecture that couples temporal alignment with entropy-guided semantics to improve long-video reasoning. The framework contributes two main mechanisms: \emph{(i)} a time-decay retrieval module that injects explicit temporal offsets into the similarity computation, thereby ranking text queries according to their true multimedia context; and \emph{(ii)} an entropy-weighted key-frame sampler that selects evenly spaced, information-dense frames, reducing redundancy while preserving representativeness. By weaving these temporal and semantic signals together, TV-RAG realises a dual-level reasoning routine that can be grafted onto any LVLM without re-training or fine-tuning. The resulting system offers a lightweight, budget-friendly upgrade path and consistently surpasses most leading baselines across established long-video benchmarks such as Video-MME, MLVU, and LongVideoBench, confirming the effectiveness of our model. The code can be found at this https URL.
Abstract (translated)
大型视频语言模型（LVLMs）作为多媒体AI研究的焦点迅速崛起。然而，面对长视频时，这些模型遇到了困难：它们的时间窗口狭窄，无法捕捉到长时间内细微的语义变化。此外，主流基于文本的检索流程主要依赖于表面级别的词汇重叠，忽视了视觉、音频和字幕通道之间的丰富时间相互依存关系。为了缓解这些问题，我们提出了TV-RAG架构，这是一种无需训练的方法，结合了时序对齐与熵引导语义来提升长视频推理能力。该框架贡献了两种主要机制：（i）一种时间衰减检索模块，在相似度计算中注入明确的时间偏移量，从而根据真实的多媒体背景排列文本查询；（ii）一种基于熵加权的关键帧采样器，选择均匀间隔且信息密集的帧，减少了冗余同时保持代表性。通过将这些时序和语义信号编织在一起，TV-RAG实现了双层推理流程，可以在不重新训练或微调的情况下与任何LVLM集成。由此产生的系统提供了一种轻量级、预算友好的升级途径，并在Video-MME、MLVU 和 LongVideoBench 等公认的长视频基准测试中持续超越大多数领先基线，证明了我们模型的有效性。该代码可在[此链接](https://this-is-the-url-you-need-to-replace.com)找到。
URL
https://arxiv.org/abs/2512.23483
