EEweb Home     ::     Graduate Courses     ::     Undergraduate Courses     ::     My Home

Course Description Form

Course number and title: EE210A Adaptation and Learning
Credits: 4
Instructor(s)-in-charge: Ali H. Sayed (
Course type: Lecture
Required or Elective: A Signals and Systems course.
Course Schedule:
Lecture: 4 hrs/week
Outside Study: 8 hrs/week
Office Hours: 2 hrs
Course Assessment:
Homework: 6 homework assignments
Exams: 1 midterm and 1 final
Grading Policy: Typically 20% homework, 30% midterm, 50% final.
Course Prerequisites: Recommended courses EE205A (Matrix Analysis) and EE241A (Stochastic Processes) or equivalent or consent of instructor. Prior training in probability theory, random processes and linear algebra is expected.
Catalog Description: Lecture, four hours; outside study, eight hours. Prior training in probability theory, random processes and linear algebra is expected. Recommended courses EE205A and EE241A or equivalent or consent of instructor. Mean-square-error estimation and filters, least-squares estimation and filters, steepest-descent algorithms, stochastic-gradient algorithms, convergence, stability, tracking, and performance, algorithms for adaptation and learning, adaptive filters, learning and classification, optimization. Letter grading.  
Textbook and any related course material:
A. H. Sayed, Adaptive Filters, Wiley, NJ, 2008 (main text).
A. H. Sayed, Fundamentals of Adaptive Filtering, Wiley, NJ, 2003.
Course Website
Additional Course Website
Topics covered in the course:
Complex gradients and complex Hessian matrices. Convex functions. Basic linear algebra and matrix theory results.
Mean-square-error estimation. Optimal estimators. Linear estimators. Linear models. Channel estimation. Channel equalization.
Kalman Filters and Wiener Filters.
Steepest-descent algorithms. Stochastic-gradient algorithms. Adaptive filters. LMS-type filters.
Mean-square error and tracking performance of adaptive filters.
Mean-square error stability and transient analysis of adaptive filters.
Least-squares designs. Least-squares adaptive filters. RLS-type filters.
Single-agent optimization, adaptation, and learning. Steepest descent methods. Stochastic gradient methods.
Adaptive implementations with decaying and constant step-sizes.
Will this course involve computer assignments? YES Will this course have TA(s) when it is offered? NO

:: Last modified: February 2013 by J. Lin ::

Copyright © 2003 UCLA Electrical Engineering Department. All rights reserved.
Please contact for comments or questions for the website.