Lecturers
Rules
- The class begins promptly, and once 10 minutes have passed, students are not permitted to enter the classroom.
- Taking pictures or recording the whiteboard is not allowed.
- Lab teams and schedules are generated randomly, and modifications are not allowed.
- To pass the course, students must pass at least 50% of the exams. The final grade is calculated based on 40% for theory and 60% for practical skills.
- During the regulation term, students can improve their grades. However, once grades are finalized, they cannot be changed.
- During the normal term, exams will be multiple choice type. Thus, the regulation term exams will be open answer.
- Lab reports must be typed by the students. If the student uses Internet information, the info must be properly referenced.
- The lab reports will be checked for plagiarism. If the report contains more than 30% similarity to other sources, the grade will be given a zero.
- The use of AI tools is prohibited. If a student is found to have used AI tools, their lab report grade will be reduced to zero.
- Grades for lab reports are determined based on the lab report rubric.
Introduction
The lecture is focused in four main topics: linear regressors and logistic classifiers, artificial neural networks, fuzzy logic and control applications. The lecture requires previous concepts of programming, control and microcontrollers. Also, is highly recommended to use a GNU/Linux distribution to follow closely several videos and tutorials in the lecture; Ubuntu distribution is a good choice.
To follow the lecture announcemnets, tutorials, and guides, please subscribe to the newsletter list:
1. Artificial intelligence fundamentals
This first section will introduce the basic concepts of linear regressors and classifiers with simple examples to relate the error with instances. Also, some basic skills will be developed to define a GNU/Linux work-flow to work with Python, Java Script (JS), Visual Studio Code, and Jupyter Notebook.
1.1 GNU/Linux environment and Python
The next link is a tutorial to install a GNU/Linux system. The procedure can be used to install any distribution (Arch, Ubuntu, Debian). If you have doubts please watch some youtube videos.1.2 Linear regressors
This first section will introduce the basic concepts of linear regressors and classifiers with simple examples to relate the error with instances. Also, some basic skills will be developed to define a GNU/Linux work-flow to work with Python, Java Script (JS), Visual Studio Code, and Jupyter Notebook.
- BeamerMachine Learning Demistified (beamer presentation)
- BeamerPredicting machine lecture
- ExampleOrdinary Least Squares method
2 Linear models: Linear regressor and logistic regressors (classifiers)
In order to improve classification tasks, the logistic regressor function is introduced as a more sophisticated method. Probability will be utilized to discriminate instances and predict new ones, with the assistance of techniques such as elastic net and Ridge regression.
- BeamerTrining Linear Models
- BeamerGradient Descent Method
- BeamerLogistic regressor
3. Artifial Neural Networks
In this section, we will be discussing the primary computational unit of an Artificial Neural Network, which is known as the Perceptron model. Our focus will be on exploring this processing unit as a basic model that utilizes the sigmoid (logistic) activation function.
-- Assigments and laboratory sessions --
- Overleaf Teplate LaTeX Engineering report template
- Grading Laboratory reports grading table
- Repository MNIST database for ANNs session