Stanford AI 课程地图 EN GitHub ↗
← 返回课程地图
Stanford · 2026—27 CS 224W 基础核心 ★★★☆☆

核心课程 · 基础核心

图机器学习

Machine Learning with Graphs

图机器学习:GNN、知识图谱、影响力最大化,把关系数据变成可学习的表示。

3–4 学分 2026 秋 Letter or Credit/No Credit ★★★☆☆
图GNN表示学习

课程定位

CS 224W 由 Jure Leskovec 讲授,是图神经网络方向最受欢迎的课程之一。它覆盖图结构分析、节点/图表示学习、GNN、知识图谱推理、传播与影响力,应用横跨社交网络、推荐、生物网络和欺诈检测。

主要内容

  1. 图基础:中心性、社区发现、幂律、图生成
  2. 节点嵌入:DeepWalk、node2vec、矩阵分解
  3. GNN:GCN、GraphSAGE、GAT、消息传递
  4. 知识图谱:TransE、RotatE、查询与推理
  5. 传播与影响力:独立级联、线性阈值、影响力最大化
  6. 图生成与自监督:图对比学习、扩散
  7. 项目:通常用真实图数据做预测/分析

课程大纲 / 课件

#主题 / 讲者
01Introduction
02Node embeddings
03Graph neural networks
04A general perspective on GNNs
05GNN augmentation and training
06Theory of GNNs
07Designing powerful graph encoders
08Graph Transformers
09Heterogeneous graphs
10Knowledge graphs
11GNNs for recommender systems
12Relational Deep Learning
13Advanced architectures in RDL
14Advanced topics in GNNs
15Towards Foundation Models for Knowledge Graphs
16Frontiers in Graph AI
17Frontiers in Graph AI II

Fall 2026 schedule;slides 1–12 已直接链接,13–17 见课程 schedule。

先修

CS 109;一门入门机器学习课程。

考核与工作量

作业 + 项目;编程以 PyTorch/PyG 为主。具体以当季官网为准。

适合谁

社交网络、推荐、风控、生物网络、知识图谱方向;想补“非欧几里得数据”建模的人。

学习建议

先理解消息传递范式,再学各种 GNN 变体。项目可以结合 CS 246 的大规模数据方法。

相关课程

官方课程描述(英文原文)

Many complex data can be represented as a graph of relationships between objects. Such networks are a fundamental tool for modeling complex social, technological, and biological systems. This course focuses on the computational, algorithmic, and modeling challenges specific to the analysis of massive graphs. By means of studying the underlying graph structure and its features, students are introduced to machine learning techniques and data mining tools apt to reveal insights on a variety of networks. Topics include: representation learning and Graph Neural Networks; algorithms for the World Wide Web; reasoning over Knowledge Graphs; influence maximization; disease outbreak detection, social network analysis. Prerequisites: CS109, any introductory course in Machine Learning.