EE 378B – Inference, Estimation, and Information Processing

Andrea Montanari, Stanford University, Spring 2019
 

A large number of problems in machine learning and signal processing require to extract information from matrices or tensors. Specific topics include:

  • Clustering;

  • Matrix completion;

  • Graph localization;

  • Dimensionality reduction and manifold learning;

  • Word2vec;

  • Embeddings;

  • Kernel regression and kernel PCA;

  • Random features implementations of kernel methods, connections to neural networks;

We will discuss general mathematical and algorithmic tools to address these problems: spectral algorithms, semidefinite programming relaxations, non-convex optimization, matrix concentration inequalities.

Class Times and Locations

  • Building 60, Room 109

  • Mon-Wed, 1:30-2:50pm

Announcement

We will have two extra sessions:

  • Friday May 17, 1:30PM-2:50PM, Room 380-380D

  • Friday May 24, 11:00PM-12:20PM, Hewlett 103

These times will be used to discuss complementary material (attendance is not required, and this material is not required for the final).