The course aims to introduce students to the field of artificial intelligence starting from the study of basic algorithms in their intimate nature. After an initial overview of the state of the art, the student will be guided in the low-level study of the functioning of machine learning algorithms to then be able to develop prediction mechanisms even for higher levels of abstraction. At the end of the course, the student will be able to analyse, design and create an Artificial Intelligence system applied to a specific problem.
teacher profile teaching materials
Non-linear optimization: the problem of local minima
Preparation of data for correct training of the neural system
Supervised training: backpropagation algorithm for calculating the gradient of the error function of an MLP
Creation of the backpropagation algorithm in C/C++ from scratch
Gradient Descent training algorithm
Implementation of the Gradient Descent algorithm in C/C++ from scratch
Stochastic Gradient Descent training algorithm
Implementation of the Stochastic Gradient Descent algorithm in C/C++ from scratch
Levenberg-Marquardt training algorithm
Implementation of the Levenberg-Marquardt algorithm in C/C++ from scratch
Training with genetic algorithms and swarm intelligence
Unsupervised training
Reinforcement training
https://d2l.ai/
Mutuazione: 20810549-2 DESIGN OF LEARNING ALGORITHMS in Ingegneria delle Telecomunicazioni LM-27 RIGANTI FULGINEI FRANCESCO
Programme
Introduction to optimization theoryNon-linear optimization: the problem of local minima
Preparation of data for correct training of the neural system
Supervised training: backpropagation algorithm for calculating the gradient of the error function of an MLP
Creation of the backpropagation algorithm in C/C++ from scratch
Gradient Descent training algorithm
Implementation of the Gradient Descent algorithm in C/C++ from scratch
Stochastic Gradient Descent training algorithm
Implementation of the Stochastic Gradient Descent algorithm in C/C++ from scratch
Levenberg-Marquardt training algorithm
Implementation of the Levenberg-Marquardt algorithm in C/C++ from scratch
Training with genetic algorithms and swarm intelligence
Unsupervised training
Reinforcement training
Core Documentation
Dive into deep learninghttps://d2l.ai/
Attendance
Attendance is optional but highly recommended.Type of evaluation
Design and implementation of a specific training algorithm/system