/
划分小节，所以我需要分析原文结构，找出可以作为小标题的部分。原文第一句是背景，接下来是问题陈述，然后是方法，最后是实验和结果。 我可以把每个逻辑段落用
包裹，然后用
来分小节。比如背景可以作为第一段，问题作为第二段，方法作为第三段，实验和结果作为第四段。但用户提到长文用
/
划分，可能需要更细致的分法。 原文结构大致是：背景、问题、方法、实验、结果。我可以用
作为主标题，然后每个部分用
。比如：
研究背景与现状
...
现有问题
...
方法框架
...
实验与结果
...
但原文中有些句子是连续的，可能需要拆分。比如第一句是背景，第二句是之前研究的作用，第三句是问题，第四句开始是方法。我需要确保每个自然段落用包裹，长文用
/
划分小节。 另外，用户要求保留所有信息，不得增删改内容，所以不能修改原文内容，只能调整结构。 所以，我应该把原文拆分成几个自然段，每个用
包裹，然后根据内容长度和结构用
/
划分。 现在开始拆分： 第一段：Multilingual automatic speech recognition (ASR) has achieved rapid progress with large-scale pretrained models and multilingual speech corpora. Previous studies have proved the pivotal role of data sampling in multilingual ASR performance. However, existing sampling strategies largely rely on time-consuming manual, heuristic-based designs. 这里可以作为研究背景和现有问题，用
或
。可能用
作为小标题，比如“研究背景与现有问题”。 第二段：To address this, we propose a linguistic-aware agentic data sampling framework that introduces a data sampling LLM agent into the multilingual ASR training loop. The agent is driven by a linguistically grounded rubric that analyzes ASR errors from phonological, morphological, and syntactic perspectives, and uses the resulting feedback together with training statistics, budget constraints, and algorithm profiles to select domain-level, sample-level, and token-level sampling strategies. This design turns evaluation results into adaptive data-selection decisions rather than treating WER/CER as final endpoints only. 这部分讲方法，可以用
作为“方法框架”。 第三段：We conduct experiments on twelve Common Voice languages using Whisper-medium as the ASR backbone. The results show that our method achieves performance better than full-data training while using less training data and lower computational resources, and substantially accelerates training convergence compared to conventional sampling strategies. 这部分是实验和结果，用
作为“实验与结果”。 现在，我需要把这些部分用
包裹，并加上适当的
标题。注意，用户要求只输出HTML片段，不要Markdown围栏，不要解释，所以直接输出HTML结构。 所以最终的HTML结构应该是：
研究背景与现有问题
Multilingual automatic speech recognition (ASR) has achieved rapid progress with large-scale pretrained models and multilingual speech corpora. Previous studies have proved the pivotal role of data sampling in multilingual ASR performance. However, existing sampling strategies largely rely on time-consuming manual, heuristic-based designs.
方法框架
To address this, we propose a linguistic-aware agentic data sampling framework that introduces a data sampling LLM agent into the multilingual ASR training loop. The agent is driven by a linguistically grounded rubric that analyzes ASR errors from phonological, morphological, and syntactic perspectives, and uses the resulting feedback together with training statistics, budget constraints, and algorithm profiles to select domain-level, sample-level, and token-level sampling strategies. This design turns evaluation results into adaptive data-selection decisions rather than treating WER/CER as final endpoints only.
实验与结果
We conduct experiments on twelve Common Voice languages using Whisper-medium as the ASR backbone. The results show that our method achieves performance better than full-data training while using less training data and lower computational resources, and substantially accelerates training convergence compared to conventional sampling strategies.
但用户提到长文用/
划分小节，可能这里需要更明确的小节标题。比如“研究背景”、“现有问题”、“方法框架”、“实验与结果”等，但原文中并没有明显的标题，所以需要根据内容合理划分。 不过，用户要求“保留原文所有信息，不得增删改内容”，所以不能添加新标题，只能根据原文结构划分。原文中没有明确标题，所以可能需要用
作为主标题，然后分小节。 或者，可能不需要
，因为原文只有几个段落，可能直接用
包裹每个自然段，然后用
来分小节。比如：
Multilingual automatic speech recognition (ASR) has achieved rapid progress with large-scale pretrained models and multilingual speech corpora. Previous studies have proved the pivotal role of data sampling in multilingual ASR performance. However, existing sampling strategies largely rely on time-consuming manual, heuristic-based designs.
方法框架
To address this, we propose a linguistic-aware agentic data sampling framework that introduces a data sampling LLM agent into the multilingual ASR training loop. The agent is driven by a linguistically grounded rubric that analyzes ASR errors from phonological, morphological, and syntactic perspectives, and uses the resulting feedback together with training statistics, budget constraints, and algorithm profiles to select domain-level, sample-level, and token-level sampling strategies. This design turns evaluation results into adaptive data-selection decisions rather than treating WER/CER as final endpoints only.
实验与结果
We conduct experiments on twelve Common Voice languages using Whisper-medium as the ASR backbone. The results show that our method achieves performance better than full-data training while using less training data and lower computational resources, and substantially accelerates training convergence compared to conventional sampling strategies.
但这样是否符合用户要求？用户说“长文用/
划分小节”，这里三个段落，可能不算长文，但需要按照结构划分。 另外，用户提到“每个自然段落用
包裹”，所以每个自然段都要用
，而小节标题用
/
。 可能正确的方式是，将原文分成几个自然段，每个用
，然后用
作为主标题，
作为小节标题。例如：
多语言自动语音识别研究进展
...
现有数据采样策略的问题
...
语言感知代理式数据采样框架
...
实验验证与结果分析
...
但原文中没有这些标题，所以不能添加，只能根据内容隐含的结构。 可能更好的方式是，将每个逻辑段落用包裹，然后用
来分小节，标题内容来自原文中的关键词。例如：
研究背景
Multilingual automatic speech recognition (ASR) has achieved rapid progress with large-scale pretrained models and multilingual speech corpora. Previous studies have proved the pivotal role of data sampling in multilingual ASR performance.
现有问题
However, existing sampling strategies largely rely on time-consuming manual, heuristic-based designs.
方法框架
To address this, we propose a linguistic