arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.02482cs.CV
收藏

X-Stream: 探索多模态大语言模型作为多流理解的多路复用器

X-Stream: Exploring MLLMs as Multiplexers for Multi-Stream Understanding

MMLab, Chinese University of Hong Kong(中大香港人工智能实验室);Huawei Inc.(华为公司)

Peiwen Sun, Xudong Lu, Huadai Liu, Yang Bo, Dongming Wu, Huankang Guan, Minghong Cai, Jinpeng Chen, Xintong Guo, Shuhan Li, Fang Liu, Rui Liu, Xiangyu Yue

AI总结:

为解决多流视频理解评估缺失的问题,提出首个基准X-Stream,包含4220个QA对和932个视频,覆盖多窗口、多视角和多设备场景,并基于信号多路复用理论评估MLLM作为多路复用器的性能,发现现有模型在并发流上仅达约50%分数。

摘要:

尽管视频流理解取得了显著进展,但实际应用(如体育直播、自动驾驶和多屏协作)本质上需要连续的多流交互。然而,现有基准局限于单流范式,在评估在线跨流推理方面存在关键空白。为填补这一空白,我们引入了X-Stream,这是首个专门用于多流流式理解的基准。X-Stream包含932个视频中精心整理的4220个QA对,评估了跨多窗口、多视角和多设备场景的11个子任务。关键的是,我们的数据集使用一种新颖的双重验证流水线构建,防止对单一流的过度依赖。此外,我们开创性地将多模态大语言模型(MLLM)概念化为朴素多路复用器,通过信号多路复用理论的视角系统评估其性能。我们广泛的在线推理实验揭示了一个严峻的现实:最先进的MLLM在并发流上表现困难,仅达到约50%的分数,且主动能力差。最终,X-Stream暴露了当前多路复用方案的权衡,为下一代多流智能体提供了实用的评估协议和经验指导。

英文摘要:

While video streaming understanding has made significant strides, real-world applications, such as live sports broadcasting, autonomous driving, and multi-screen collaboration, inherently demand continuous, multi-stream interactions. However, existing benchmarks are confined to single-stream paradigms, leaving a critical gap in evaluating online, cross-stream reasoning. To bridge this, we introduce X-Stream, the first benchmark dedicated to multi-stream streaming understanding. Comprising 4,220 rigorously curated QA pairs across 932 videos, X-Stream evaluates 11 subtasks across multi-window, multi-view, and multi-device scenarios. Crucially, our dataset is constructed using a novel dual-verification pipeline that prevents over-reliance on a single stream. Furthermore, we pioneer the conceptualization of multi-modal large language models (MLLMs) as naive multiplexers, systematically evaluating their performance through the lens of Signal Multiplexing Theory. Our extensive online inference experiments reveal a stark reality: state-of-the-art MLLMs struggle significantly with concurrent streams, achieving only about 50% score and exhibiting poor proactive ability. Ultimately, X-Stream exposes the trade-off of current multiplexing schemes, providing both a practical evaluation protocol and empirical guidance for next-generation multi-stream agents.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑