---
格式版本: 2
标题: "How smoothing the affinity matrix affects neighborhood preservation in t-SNE"
原文链接: "https://arxiv.org/abs/2608.17190"
发布日期: "2026-08-17"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Mon, 17 Aug 2026 23:05:51 UTC (5,596 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-08-20T15:43:26+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-08-20T15:39:49+08:00"
入库时间: "2026-08-20T07:43:26.737Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=bandwidth&searchtype=all"
匹配关键词:
  - "bandwidth"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 15
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该资料为arXiv机器学习论文，主题为t-SNE降维算法，与超节点/AI Rack/机柜级AI基础设施完全无关，无任何相关技术、商业或部署信息。"
AI质检模型: "ali-deepseek-v4-flash"
AI质检时间: "2026-08-20T15:49:02+08:00"
AI主题相关性: 0
AI来源权威性: 8
AI新颖性: 2
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 5
AI摘要: "该研究探讨t-SNE中亲和矩阵概率分布锐度对不同尺度邻域保留的影响，并提出由参数gamma控制的行方向幂变换，可平滑或锐化各行且保持稀疏性与秩序。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-07T03:24:50.385Z"
采集批次: "2026年8月20日14点19分32秒"
采集批次ID: "20260820-141932-079"
去重键: "https://arxiv.org/abs/2608.17190"
---

## Computer Science > Machine Learning

## Title:How smoothing the affinity matrix affects neighborhood preservation in t-SNE

Authors:[Shirin Mohebi](https://arxiv.org/search/cs?searchtype=author&query=Mohebi,+S), [Guillaume Bied](https://arxiv.org/search/cs?searchtype=author&query=Bied,+G), [Jefrey Lijffijt](https://arxiv.org/search/cs?searchtype=author&query=Lijffijt,+J)

[View PDF](https://arxiv.org/pdf/2608.17190) [HTML (experimental)](https://arxiv.org/html/2608.17190v1)

> Abstract:Dimensionality reduction methods are instrumental to visualize high-dimensional data, and t-SNE stands as one of the most widely used methods due to its emphasis on local neighborhood preservation. A central component of t-SNE is the affinity matrix, which expresses pairwise similarities in the form of symmetrized probabilities, over which the optimization problem of t-SNE is defined. We study how the sharpness of this probability distribution affects neighborhood preservation at different scales. We introduce a row-wise power transform controlled by a parameter gamma that can smooth or sharpen each row of the affinity matrix while preserving sparsity and rank order. We show that this transform is equivalent to rescaling the Gaussian bandwidth and thus to changing the perplexity. However, as the sharpness of the probability distribution varies per point, a fixed gamma leads to point-dependent effective perplexities, making it distinct from changing the global perplexity. Empirically, we find that sharpening improves preservation of the very nearest neighbors, while smoothing improves preservation of broader local neighborhoods, outperforming alternative affinity constructions including multiscale methods in the mid-local range.

| Comments: |  |
| --- | --- |
| Subjects: | Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | [arXiv:2608.17190](https://arxiv.org/abs/2608.17190) \[cs.LG\] |
|  | (or [arXiv:2608.17190v1](https://arxiv.org/abs/2608.17190v1) \[cs.LG\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.17190](https://doi.org/10.48550/arXiv.2608.17190) |

## Submission history

From: Shirin Mohebi \[[view email](https://arxiv.org/show-email/87620b49/2608.17190)\]  
**\[v1\]** Mon, 17 Aug 2026 23:05:51 UTC (5,596 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.17190) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))