EE 378A: Statistical Signal Processing


  • Welcome to the EE378a course!

  • We will use Coursework for grade records. Please go to Coursework to look up your homework grades.

  • We have set up a forum in Piazza. You can sign up as a student here.

Course Overview

This year's EE378A offering has been substantially revised to include recent research results in data science, as well as a comparison between and analysis of existing frameworks for data analysis and processing. Among other questions, we will consider the following:

  • What is the difference between statistical decision theory and learning theory?

  • What is the difference between statistics and machine learning?

  • What are the bias-variance trade-offs in decision theory and in learning theory?

  • How can we incorporate prior knowledge in real world problems, and derive effective algorithms?

  • Can we make inference on data without prior assumptions on the data generation mechanisms?

We will also analyze algorithms that perform well in one theoretical framework but are highly sub-optimal in another, and contrast with algorithms that perform well simultaneously under multiple frameworks. Students will be exposed to a variety of disciplines, and connections between them.


The course will cover and emphasize the importance of both theory and algorithms. Grading will be based primarily on about 4 homework sets, and a final project. Project topics will be offered, but students are welcome to propose their own project topics.