Estimation, Detection, and Learning (EE 60573)

University of Notre Dame


Spring 2027 - Enroll in EE 60573 section 01
place/data TBD

Course Vault, UNDER CONSTRUCTION

Description: This course studies methods for parameter estimation and signal detection, in engineering and machine learning applications. The course has three modules. The first module (Estimation Theory) studies the Cramer-Rao Bound, sufficient statistics, Minimum Variance Estimators, Maximum Likelihood Estimators, Bayesian Least Squares Estimators, and the discrete-time Kalman filter. The second module (Detection Theory) studies Neyman-Pearson and Baysian detectors, matched filters, and composite hypothesis testing. The third module (Learning Theory) examines PAC learning and VC-methods used to bound a trained model's generalization performance.
Topics/Lecture
  • Estimation Theory
    Cramer-Rao Bound, Sufficient Statistics and Minimum Variance Unbiased Estimators, Maximum Likelhood Estimators, Bayesian Least Squares Estimators, Discrete-time Kalman Filters
  • Detection Theory
    Neyman-Pearson Detectors, Bayes Detectors, Matched Filters, Composite Hypothesis Testing
  • Learning Theory
    Supervised Learning, PAC Learning, VC-theory and model generalization
References:
  • Steven Kay, "Fundamentals of Statistical Signal Processing (volumes 1 and 2)", Prentice Hall, 1993
  • V. Poor, "An Introduction to Signal Detection and Estimation", Springer, 2nd edition, 1994
  • T. Kailath, "Lecture on Linear Least-Squares Estimation", Springer, 1976
  • V.N. Vapnik, "Statistical Learning Theory", Wiley, 1998
Grading: Homework 10% - 2 Midterm Exams 60% - Final Exam 30%
Instructor: Michael Lemmon, Dept. of Electrical Engineering, University of Notre Dame, Fitz 266