Intelligence Algorithms for Data Analysis 2021-II

Module: Machine Learning 'n' Deep Learning

Topics

  • Introduction

  • Machine Learning and its applications with Scikit-Learn

  • Artificial Neural Networks, Deep Learning and its applications with Pytorch

Link for meeting (or virtual lectures: Nov 29 (4h) and 30 (4h), Dec 1 (4h), 2 (4h), 4 (8h), 6 (4h), 7 (4h), 9 (4h), 11 (8h), 13 (4h), 14 (4h), and 15 (4h) - 56 hours -)

Final GRADES

Google meet: https://meet.google.com/tjt-soxd-zgo

Video recorded lectures (You can watch them as long as you have an institutional email account, otherwise you must ask permission):

Evaluation

  • Project (50 %): You ought to apply what you've learned in this course. You can use a dataset from either the UC Irvine Machine Learning Repository or Kaggle (you can find the links below, in resources section). You ought to share your notebooks with me through the following Google Form (open it as long as you're ready to submit your project notebook, otherwise, don't do so). You might work solo or team up with another two mates, if you choose the latter option, each member of the team must present as clear as possible their contribution. If the notebook don't run (for instance because you've not included the commands for downloading the dataset) you might get the minimum grade (zero). The deadline is due December 21th.

  • Coursework (50%): assignments, tests, quizzes, and so forth. Each course work weighs the same.

Attendance record:

Coursework:

  • Assignment 01 (submit the url of your notebook here / enviar la url de sus cuadernos aquí)

  • Assignment 02 - testing several values of the regularization parameter (The submission deadline is due December 6th)

  • Assignment 03 - testing several kernels and regularization parameters with Kernel Ridge Regression (The deadline is due December 8th)

  • Assignment 04 - testing polynomial kernel with Pytorch (The deadline is due December 7th at noon)

  • Assignment 05 - testing weigh decay on artificial neural networks for regression (The deadline is due December 11th 13th)

  • Assignment 06 - you ought to test the source code that implements the artificial neural network presented in lectures, adopting both Scikit-Learn and Pytorch. In this assignment you ought to share with me the following notebooks: 1) Using ReLU activation function in the hidden layer neurons with Scikit-Learn, 2) Using Sigmoid (a.k.a., Logistic) activation function in the hidden layer neurons with Scikit-Learn, 3) Using tanh activation function in the hidden layer neurons with Scikit-Learn, 4) Using ReLU activation function in the hidden layer neurons with Pytorch, 5) Using leaky ReLU activation function in the hidden layer neurons with Pytorch, 6) Using Sigmoid activation function in the hidden layer neurons with Pytorch, 7) Using tanh activation function in the hidden layer neurons with Pytorch. The deadline is due December 17th.

  • Project - Read carefully the above presented instruction before submitting. The deadline is due December 21th.

Datasets used in the course:

Resources: