课程定位
CS 229M(STATS 214)是数学味最重的一门 ML 课程,面向想做 theory 的研究生。它不教你调模型,而是训练你用集中不等式、一致收敛、优化理论去分析现代大规模非线性模型。
主要内容
- 集中不等式:Hoeffding、Chernoff、McDiarmid
- 泛化理论:一致收敛、VC 维、Rademacher 复杂度
- 非凸优化:梯度下降、鞍点、landscape
- 隐式正则:过参数化、双下降、边缘稳定
- 无监督与域适应:表示学习理论
- 现代模型分析:Transformer/深度网络的近似与泛化
先修
MATH 51;STATS 117;CS 229 或 STATS 315A。
考核与工作量
作业 + 期末/项目,以证明和论文阅读为主。具体以当季官网为准。
适合谁
想做 ML theory、读博、或需要严谨分析能力的研究者;也适合数学/统计背景转 ML 的人。
学习建议
先补实分析和概率论。不要指望一次听懂,配合讲义和论文反复推。
相关课程
官方课程描述(英文原文)
How do we use mathematical thinking to design better machine learning methods? This course focuses on developing mathematical tools for answering this question. This course will cover fundamental concepts and principled algorithms in machine learning, particularly those that are related to modern large-scale non-linear models. The topics include concentration inequalities, generalization bounds via uniform convergence, non-convex optimization, implicit regularization effect in deep learning, and unsupervised learning and domain adaptations. Prerequisites: MATH 51 and STATS 117 and either CS 229 or STATS 315A. See https://statistics.stanford.edu/course-equiv for equivalent courses in other departments that satisfy these prerequisites.