# X-Stream： 探索MLLM作为多路复用器的多流理解能力

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-06-01 08:00
- AIHOT 分数：64
- AIHOT 链接：https://aihot.virxact.com/items/cmpw15rrx03flslukdjxyt336
- 原文链接：https://arxiv.org/abs/2606.02482

## AI 摘要

专注于多流视频流理解的基准测试X-Stream被提出，包含4220个QA对、932个视频和11个子任务，覆盖多窗口、多视图和多设备场景。研究首次将多模态大语言模型（MLLM）的概念化为信号复用器，并基于信号复用理论进行评估。在线推理实验显示，当前最先进的MLLM在处理并发视频流时表现挣扎，得分仅约50%且主动能力较差。该基准揭示了现有复用方案的权衡，为多流智能体提供了评估协议和实证指导。

## 正文

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.