1.
明确角色和目标 * 我是专业的新闻排版工程师。 * 我的任务是将给定的纯文本新闻正文排版成语义化的HTML片段。 2.
拆解约束与规则 * 规则1：仅输出HTML片段本身。不要Markdown代码围栏，不要解释。 * 规则2：每个自然段落用 `
` 包裹。长文使用 `
` 或 `
` 划分小节。 * 规则3：列表内容使用 `
- `，引用使用 `
`，代码使用 `
`，重要句子使用 ``。
* 规则4：保留所有原始信息，不得增删改内容。
* 规则5：不要输出文档级标签，如 ``, ``, `` 等。
3. 分析源文本
* 文本是一段学术摘要。
* 它讲述：
* 现有时空图卷积网络的局限性（固定大小的邻接矩阵、传感器数量N、对未见传感器无输出）。
* 提出OmniST：拓扑不变的预测器，具有归纳拆分保形校准，在任意位置提供无分布且覆盖保证的预测区间。
* 拓扑不变性的机制：基于坐标的标记化、每次前向传播从地理坐标中进行k近邻注意力、通过双向GRU进行逐节点时间处理。
* 在德里空气质量数据上的评估（37个站点，2019-2023），30天PM2.5 MAE为31.0 ug/m3，在移除40%的站点后最多下降1.1%。
* 在未见站点上的归纳预测优于最近邻插值（约12%）和学习的克里金GNN（IGNNK，约21%），并且具有有限样本覆盖保证。
* 在4个真实数据集（德里和北京的空气质量、NOAA GHCN天气、Caltrans PeMSD7-M交通）中，当节点丢弃率为30-50%时，90%或以上的保证仍然成立，并且在325个真实传感器（完整的PeMS-BAY）上依然成立。
* 修复时间漂移失败：静态保形会随时间降低覆盖率（后期约0.50），但滚动自适应保形估计器恢复了目标覆盖率（总体0.89，后期0.96），代价是区间更宽。
* 文本是一段连续的英文段落。我需要决定如何组织HTML结构。
* 由于它是标准的摘要，用一个 `` 包裹整个文本从技术上讲是可以接受的。然而，规则3提到了“长文用
/
划分小节”。这是摘要还是一篇长文？提示词要求排版“新闻正文”，它通常具有清晰的结构。即使它是一段文本，我也可以识别关键主题作为小节，或者直接使用 `
` 包裹。
* 仔细重读：“每个自然段落用
包裹”。提供的文本只是一个自然段落。如果将其拆分为多个段落，就会违反“不得增删改内容”。因此，我必须将其保留为单个自然段落。
* 为了使其更具语义化，我可以为关键概念添加 ``。我可以使用 `
` 来介绍主要思想，但标题可能被视为“增加内容”。规则4说“不得增删改内容”。所以，我绝对不能添加标题或节标签，因为它们不在原文中。
* 等等，规则2说：“长文用/
划分小节”。这意味着我*可以*使用它们，但我*不能*增加内容。如果我添加 `
Introduction
`，我就增加了文本。所以除非源文本有标题，否则我不应该添加标题。源文本只有一段。
* 因此，我*只能*对已有的段落使用 ``。
* 让我们检查一下是否可以使用 ``。“重要句子可用”。我可以对关键句子进行加粗。 * 让我们检查一下是否可以使用列表/块引用。原文中没有列表或块引用。 * 所以HTML结构将仅仅是一个包裹长段落的 `
` 标签，并在内部使用 `` 标签来强调关键术语/句子。 4. 草拟HTML * 原始文本：“Standard spatiotemporal graph convolutional networks need a fixed-size adjacency matrix, coupling model weights to a specific sensor count N; they cannot produce any output at a sensor absent from training. We present OmniST, a topology-invariant forecaster whose computation has zero structural dependency on the number of nodes, and which, composed with inductive split-conformal calibration, gives distribution-free, coverage-guaranteed prediction intervals at arbitrary (including never-seen) sensor locations, preserving that guarantee as sensors are added or removed. Topology invariance follows from coordinate-based tokenization, k-nearest-neighbor attention from geographic coordinates at each forward pass, and per-node temporal processing via a bidirectional GRU. On real Delhi air-quality data (37 stations, 2019-2023), OmniST attains a 30-day PM2.5 MAE of 31.0 ug/m3 and degrades at most 1.1 percent when 40 percent of stations are removed. Its inductive forecasts at unseen stations beat nearest-neighbor interpolation (about 12 percent) and a learned kriging GNN (IGNNK, about 21 percent), and, unlike either, carry a finite-sample coverage guarantee. The guarantee holds at 90 percent or above under 30-50 percent node dropout across four real datasets over three domains (Delhi and Beijing air quality, NOAA GHCN weather, Caltrans PeMSD7-M traffic), and still holds at 325 real sensors (full PeMS-BAY). We also fix a temporal-drift failure: static conformal under-covers over time (about 0.50 late), but a rolling adaptive-conformal estimator restores target coverage (0.89 overall, 0.96 late) at the cost of wider intervals.” * 我要加粗哪些句子？ * “We present OmniST, a topology-invariant forecaster whose computation has zero structural dependency on the number of nodes, and which, composed with inductive split-conformal calibration, gives distribution-free, coverage-guaranteed prediction intervals at arbitrary (including never-seen) sensor locations, preserving that guarantee as sensors are added or removed.” - 这是主要贡献。 * “Topology invariance follows from coordinate-based tokenization, k-nearest-neighbor attention from geographic coordinates at each forward pass, and per-node temporal processing via a bidirectional GRU.” - 这是方法。 * “Its inductive forecasts at unseen stations beat nearest-neighbor interpolation (about 12 percent) and a learned kriging GNN (IGNNK, about 21 percent), and, unlike either, carry a finite-sample coverage guarantee.” - 这是关键的实验结果。 * “We also fix a temporal-drift failure: static conformal under-covers over time (about 0.50 late), but a rolling adaptive-conformal estimator restores target coverage (0.89 overall, 0.96 late) at the cost of wider intervals.” - 这强调了对局限性的修复。 * 让我们构建 `
` 标签： ```html
Standard spatiotemporal graph convolutional networks need a fixed-size adjacency matrix, coupling model weights to a specific sensor count N; they cannot produce any output at a sensor absent from training. We present OmniST, a topology-invariant forecaster whose computation has zero structural dependency on the number of nodes, and which, composed with inductive split-conformal calibration, gives distribution-free, coverage-guaranteed prediction intervals at arbitrary (including never-seen) sensor locations, preserving that guarantee as sensors are added or removed. Topology invariance follows from coordinate-based tokenization, k-nearest-neighbor attention from geographic coordinates at each forward pass, and per-node temporal processing via a bidirectional GRU. On real Delhi air-quality data (37 stations, 2019-2023), OmniST attains a 30-day PM2.5 MAE of 31.0 ug/m3 and degrades at most 1.1 percent when 40 percent of stations are removed. Its inductive forecasts at unseen stations beat nearest-neighbor interpolation (about 12 percent) and a learned kriging GNN (IGNNK, about 2