Description and objectives:
In this project the student will familiarize itself with the problem of classification via the perceptron. Classical results on the capacity of the perceptron for random inputs in R^n and a simple perceptron learning algorithm wil be studied through the literature.
In a second phase the student will numerically investigate simple algorithms for simple models of structured sparse inputs.
Prerequisite:
A taste for programming
Lab and supervisor:
Nicolas Macris (for more information please contact me)
Number of students: