---
格式版本: 2
标题: "Gromov-Monge Flow Matching for Equivariant Graph Generation"
原文链接: "https://arxiv.org/abs/2608.26961"
发布日期: "2026-08-27"
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\\]** Thu, 27 Aug 2026 11:01:03 UTC (420 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: "2026-08-30T15:13:54+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: "2026-08-30T15:11:38+08:00"
入库时间: "2026-08-30T07:13:54.111Z"
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=Scale-up&searchtype=all"
匹配关键词:
  - "Scale-up"
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: 18
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "正文主线是图生成模型的Gromov-Monge流匹配方法，与超节点、AI机架及机架级基础设施无关。来源为可追溯的arXiv预印本摘要页；固定知识库未发现该论文的重复记录，但未命中不能证明首次出现，且其新增理论、训练耦合方法和生成质量结果不属于目标业务新增。正文没有机架架构、互连、供电、散热、RAS或量产部署信息，命中弱相关强否决，当前页面不值得作为超节点业务信息源。"
AI质检模型: "gpt-5.6-sol"
AI质检时间: "2026-08-30T15:14:29+08:00"
AI主题相关性: 0
AI来源权威性: 9
AI新颖性: 2
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 7
AI评分提示词版本: "v17-精简生产版"
AI评分提示词SHA256: "48fb9777f386026761b4873eaff30807694fb11e9b352d7c69bf2dfde750cc7d"
AI评分知识库版本: "knowledge_base_v1-20260819+runtime.55"
AI评分知识库SHA256: "ca56afec1ff616b1fa4394cd1d08b1e2a6af47eccbe70c802c9f1fea6037e803"
AI评分知识库检索词: "[\"Scale-up\",\"PDF\",\"arxiv.org/pdf/2608.26961\",\"HTML\",\"arxiv.org/html/2608.26961v1\",\"v1\",\"arxiv.org/abs/2608.26961\",\"arxiv.org/abs/2608.26961v1\",\"doi.org/10.48550/arXiv.2608.26961\",\"arxiv.org/show-email/747aac6b/2608.26961\",\"UTC\",\"KB\"]"
AI评分知识库命中: "[{\"id\":\"july-correct-0111\",\"title\":\"StrataCL: Fabric-Native Communication Library for Production Supernodes\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"PDF\",\"HTML\",\"v1\",\"UTC\",\"KB\"],\"rank\":-23.633258421111247},{\"id\":\"runtime-569639751a0dbe7ef3ffccc5\",\"title\":\"[2608.17503] Predict Before Replay: Joint FEC and Flight Control for Reliable Scale-Up Links\",\"sourceType\":\"ai_excellent_article\",\"time\":\"2026-08-18\",\"matchedTerms\":[\"Scale-up\",\"PDF\",\"HTML\",\"v1\",\"UTC\",\"KB\"],\"rank\":-19.058979468852325},{\"id\":\"july-correct-0010\",\"title\":\"Schneider Electric and AMD release first Helios platform reference design to accelerate AI Factory deployment\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"PDF\",\"v1\",\"KB\"],\"rank\":-10.38877158402186},{\"id\":\"july-correct-0115\",\"title\":\"锚定 300kW 整机柜演进方向 OAII 社区三项规范联合发布，树立AIDC基础设施建设标准答案\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"HTML\",\"v1\"],\"rank\":-6.992347863111062},{\"id\":\"july-correct-0062\",\"title\":\"Why AI Racks Need an Open Signal Conditioning Standard\",\"sourceType\":\"labeled_article\",\"time\":\"2026-07\",\"matchedTerms\":[\"Scale-up\",\"PDF\",\"v1\"],\"rank\":-6.929251892617623}]"
AI摘要: "该论文提出Gromov-Monge Flow Matching方法，用于等变图生成。该方法通过Gromov-Monge距离优化节点对齐构造训练耦合，兼容现有置换等变架构。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-08-30T18:08:48.323Z"
采集批次: "2026年8月30日14点11分37秒"
采集批次ID: "20260830-141137-130"
去重键: "https://arxiv.org/abs/2608.26961"
---

## Computer Science > Machine Learning

## Title:Gromov-Monge Flow Matching for Equivariant Graph Generation

Authors:[Moritz Piening](https://arxiv.org/search/cs?searchtype=author&query=Piening,+M), [Christian Wald](https://arxiv.org/search/cs?searchtype=author&query=Wald,+C)

[View PDF](https://arxiv.org/pdf/2608.26961) [HTML (experimental)](https://arxiv.org/html/2608.26961v1)

> Abstract:Graphs are invariant under node permutations, motivating the use of permutation-equivariant architectures in generative models. In flow matching, however, symmetry may also enter the source--target coupling: once graph pairs are compared up to node relabeling, the natural Wasserstein geometry is that of the graph quotient space. The Euclidean quotient metric of this space coincides with the Gromov--Monge distance, obtained by optimally relabeling the nodes. We develop this perspective theoretically, showing that quotient couplings can be lifted to aligned representatives without additional cost and that symmetrization yields equivariant flow-matching minimizers, including for categorical endpoint prediction. In practice, exact Gromov--Monge alignment is intractable, so we construct minibatch couplings using efficient Gromov--Wasserstein-type relaxations and lower bounds for the inner node alignment, optionally combined with an outer assignment between graphs. The resulting procedure changes only the training coupling and is compatible with standard permutation-equivariant architectures. Across continuous graph and categorical molecular generation, these structure-aware couplings substantially improve sample quality at small integration budgets, while our scaled-up molecular models remain competitive under conventional many-step sampling.

| Comments: |  |
| --- | --- |
| Subjects: | Machine Learning (cs.LG); Optimization and Control (math.OC); Machine Learning (stat.ML) |
| Cite as: | [arXiv:2608.26961](https://arxiv.org/abs/2608.26961) \[cs.LG\] |
|  | (or [arXiv:2608.26961v1](https://arxiv.org/abs/2608.26961v1) \[cs.LG\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2608.26961](https://doi.org/10.48550/arXiv.2608.26961) |

## Submission history

From: Moritz Piening \[[view email](https://arxiv.org/show-email/747aac6b/2608.26961)\]  
**\[v1\]** Thu, 27 Aug 2026 11:01:03 UTC (420 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2608.26961) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))