ECE 544NS is a special topics course in pattern recognition. Content varies every year, but usually includes error metrics (e.g., information-theoretic and perceptron-based) and optimization (e.g., neural network, Bayesian, stochastic, and convex programming techniques) for the supervised, semi-supervised, and unsupervised estimation of probability densities, feature selection, regression and classification. In fall 2016, the course will focus on neural networks, as a mechanism by which we may understand a wide variety of problems in pattern recognition. More detail is provided in the lists of quizzes and homework assignments.

Pre-requisites

Probability and linear algebra.

Text

Students are expected to read, understand, and apply the articles assigned as background for each homework assignment.

Grading

Homework

Each homework will include (a) a written part, in which you derive or design an algorithm, (b) a Python part, in which you code and test your solution, (c) many will also include a TensorFlow part, in which you test a solution to the same problem using TensorFlow.

Exams

There will be two exams, each consuming a full class period, and one long written homework that counts as if it were an exam. Problems will be designed to be similar to problems you've previously seen in quizzes and/or homework.

Project

During the last two weeks of the semester, in place of homework, students will design their own final project.